Bedload Sediment Transport and Channel
Morphology of a Southeast Alaskan Stream
by
Margaret A. Estep
A THESIS
submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Master of Science
Completed June 2, 1982
Commencement June 1983
AN ABSTRACT OF THE THESIS OF
Margaret A. Estep
in
for the degree of
Forest Engineering
Title:
presented on
Master of Science
June 2, 1982
BEDLOAD SEDIMENT TRASNPORT AND CHANNEL MORPHOLOGY OF A
SOUTHEAST ALA
STREAM
.
Abstract approved:
/
Qci
Paul Adams
This study was conducted at portions of Trap Bay Creek, a mediumsized third-order stream on Chichagof Island, Southeast Alaska, to
1) quantify short-term sediment transport and channel morphology
changes, 2) relate measured sediment transport rates to the major
hydrologic parameters that appeared to determine the mechanisms of
sediment transport, and 3) evaluate how bedload transport can influence
channel morphology.
Morphologic characteristics were evaluated by
means of plani-metric surveys and cross-sectional measurements made
in July and August of 1980 and August of 1981.
Bedload and suspended
sediment data were collected during..the Fall of 1980 along with data
on streamflow and precipitation.
Morphologic evaulations indicated that the stream. is a dynamic
system and that it appeared to be widening and aggrading during 198081.
Large organic debris, especially fallen trees, are important in
stream morphology, especially above the zcne of tidal influence.
tides, as well as human activity, probably contributed to recent
The
morphological changes in lower reaches of the channel.
Problems involved in processing suspended sediment samples
sulted in a limited amount of suspended sediment data.
re-
However, data
that were collected from one point on the stream indicated that suspended sediment concentrations were low, usually less than
and did not exceed 90 mg 1
interiyal storm event.
5
ingl1,
1
during an approximate 2 to 5-year return
Under average storm conditions, therefore, sus
pended sediment transport appears to be supply limited and constitutes
a small portion of total sediment transport.
Bedload sediment samples were collected from a short pool-riffle
study reach during a total of ten storms with streaf low ranging from
0.01 to 1.26 m3s1km2. Bedload sediment transport ranged from 3.9
to 4400 kgohr1, with peak transport rates occurring during peak streamflow.
Regression relationships were developed between bedload transport, gravel-sized inorganic bedload transport, coarse particulate orgariic matter transport, two particle size diameter classes (D50 and
D90), and stream discharge during the ten storms.
This analysis re-
vealed that total inorganic- bedload transport was more strongly relat-
ed to discharge than was transport of the large size category, coarse
particulate organic matter transport tended to be more strongly related
to streamflow than bedload discharge, and that neither of the particle size diameters had any consistent relationship to streaflow.
Bedload transport during the ten storms was further evaluated in
terms of the sampling sites that were used, i.e. riffles above and below a depositional area approximately 20 m in length.
Transport
tended to be greater, in ternis of amount transported, at the upper riffle for most of the storni events.
The opposite was true during the
largest storm of the season and a storm which occurred a week later.
It may be that bedload sediment is transported past the upper riffle
by lesser magnitude events axd is temporarily stored in the pool.
Transport out of the pool requires events of greater magnitude.
Sup-
ply limitations also appear to determine bedload transport in Trap
Bay Creek.
Keywords:
Helley-Smith sampler, pool-riffle sequence, armor layer,
coarse particulate organic ulatter, suspended sediment, Southeast
A.laska.
APPROVED:
Assistant Professor of Forest Hydrology in charge of major
Head of Department of Forest Engineering
Dean of Graduate School
Date thesis is presented
Typed by Jane A. Tuor for
2 June 1982
Margaret A. Estep
ACKNOWLEDGEMENTS
This study was made possible by a grant from the Pacific Northwest Range and Experiment Station through the Forestry Sciences Laboratory, Juneau, Alaska, obtained by Dr. Robert L. Beschta.
Labora-
tory facilities were made available by the Forestry Sciences Lab
and by the Department of Forest Engineering, Oregon State University,
Corvallis
I would like to thank Bob Beschta for help in all phases of the
project, including building sampling bridges.
I would also like to
thank all of the folks in Juneau from the Lab and from the Alaska
Deparent of Fish and Game - Sport Fish Division for all of their
help and encouragement.
Thanks, also, to John Adams of the Region
X Supervisor's Office, and to the hydrologists from the Sitka Ranger
Station, Tongass National Forest.
who saved me fro
Special thanks to Dick Orchard
the flood, and to Roy and Ede Sidle for providing
housing, food and fellowship.
Finally, thanks to
y coittee members
for their coents and criticisms of this manuscript, and especially
to Paul Adams for being an excellent substitute.
TABLE OF CONTENTS
Page
INTRODUCTION
Background and objectives of study
4
LITERkTIJR.E REVIEW
6
Forest management activities and fluvial transport
processes
Sediment in streams
Bedload measurement and prediction techniques
1
6
10
13
WATERSHED DESCRIPTION
Stream characteristics
27
32
NETHODS
Precipitation and streamf low
Channel morphology
Total suspended solids and turbidity
Bedload transport measurements
Bedload sample analysis
Calculation of bedload discharge
Organic matter
37
RESULTS AND DISCUSSION
Precipitation and streamf low
Bed composition
Channel morphology
Total suspended solids
Bedload discharge
Bedload transport at Flynn Creek, Oregon vs.Trap Bay,
Alaska
47
47
113
CONCLUSIONS
118
LITERATURE CITED
121
37
38
39
40
41
43
45
59
61
75
79
APPENDICES
List of common and scientific names of plants and
animals referred to in this paper
Equations for predicting mean annual flow, mean
monthly flows for August through November, and peak
flows for storms of various return periods for Trap
Bay Creek, Chichagof Island, Alsaka, taken from the
Water Resources Atlas for Alaska (1979)
126
127
TABLE OF CONTENTS - continued
Page
Relationships between bedload transport, coarse
particulate organic matter transport9 two particle
diameters, and streamflow which were not included
in the text
Morphometric characteristics of Trap Bay Creek,
Chicagof Island, Alaska
128
144
LIST OP FIGURES
Figure
1
2
3-10
11
Page
General map of the Trap Bay Watershed, Chichagof
Island, Alsaka
28
Flow chart for bedload sample analysis
47
Precipitation intensities and durations and the resultant hydrographs for the ten stornis during which bedload
48-5)
sampling was conducted
Stage-discharge rating curve for Trap Bay Creek, Chichagof Island, Alaska
60
12
Size distribution curves developed from random samples
of surface particles in the study reach, Trap Bay Creek,
63
Chichagof Island, Alaska
13
Stream width and distance of thalweg from left bank, and
thalweg depth of Trap Bay Creek, Chichagof Island, Alas64
ka
14
Thalweg profile and planimetric map of the study reach
along Trap Bay Creek, Chichagof Island, Alaska
15-18
19
65
Planimetric maps of the lower 1844 m (6050) of Trap Bay
Creek showing changes from August, 1980 to August, 1981.67-70
Net channel changes from 14 August 1980 to 21 August
1981 within the study reach of Trap Bay Creek, Chichiagof Island, Alaska
73
20
Total suspended soilds (TSS) versus stream discharge for
two time periods during the Fall of 1980, Trap Bay
78
Creek, Chichagof Island, Alaska
21
Total suspended solids, total suspended organics, and
stream discharge over time during the storni of 24-25
September 1980, Trap Bay Creek, Chichagof Island, Alaska
22
23-32
80
Total suspended solids, total suspended organics, and
stream discharge over time during the storm of 30 Sept.81
1 Oct. 1980, Trap Bay Creek, Chichagof Island, Alaska
Discharge (Q), bedload (BLD) and organic bedload
(BLDor) in transport, and average particle diameter of
bedloaa sample (D5) over time for ten storms during
which bedload sampling took place, Trap Bay Creek, Chi83-92
cagof Island, Alaska
LIST OF FIGURES
continued
Figure
33
34
35
Page
Bedload, coarse particulate organic 'atter, and
particle diameter vs. streamf low relationships for
all data collected from Trap Bay Creek, Chichagof
Island, A1aska during the Fall, 1980, storm season....
Bedload and coarse particulate organic matter vs.
streamfiow relationships for data collection from each
sampling site on Trap Bay Creek, Chichagof, Alaska,
during the Fall, 1980, storm season
99
100
Bedload and coarse particulate organic matter vs.
streamflow relationships for rising limb and falling
limb data collected from Trap Bay Creek, Chichagof Islaxd, .A.lask.a......000.0..0.000606...00000 O09000
......103
36
37
Bedload vs streamf low relationships for individual
storm events for data collected from Trap Bay Creeks
Chichagof Island, Alaska, during the Fall, 1980, storm
season
107
Coarse particulate organic matter (CPOM) vs. streamf low relationships for individual storm events for data
collected from Trap Bay Creek, Chichagof Island, Alaska, during the Fall, 1980, storm season
108
LIST OF TABLES
Table
1
2
3
4
5
6
7
8
9
10
11
Page
Percent of total load transported as bedload during
peak flows for several streams in the Pacific Northwest
12
Mean percentage of total bedload in each particle size
class (Y/X) and rate of change in percentage (B) as
bedload transport rate changes
21
Measured peak flows and estiiated recurrence intervals
for Fall, 1980, storms, Trap Bay Creek, Chichagof Island, Alaska
58
Size distributions of particles selected at random from
the armor layer of the study reach, Trap Bay Creek,
Chichagof Island, Alaska
62
Width/Depth ratios for selected stations in 1980 and
1981, Trap Bay Creek, Chichagof Island, Alaska
74
Summary of total suspended solids data collected from
Trap Bay Creek, Chichagof Island, Alaska, during the
Fall of 1980
77
Change in sediment storage computed from cross-sectional area changes from September, 1980, to August, 1981,
for Trap Bay Creek, Chichagof Island, Alaska
94
Approximate amount of sediment transported by Fall,
1980, storms, Trap Bay Creek, Chichagof Island, Alaska.
95
Relationships between bedload tranpsort, coarse particulate organic matter, two particle diameters, and
streamflow for the 1980 FallS storm season at Trap Bay
Creek, Chichagof Island, Alaska
97
Relationships between bedload transport, coarse particulate organic matter, two particle diameters, and
streamflow for each sampling site on Trap Bay Creek,
Chichagof Island, Alaska, for the Fall, 1980, storm
season
98
Relationships between bedload transport, coarse particulate organic matter, two particle diameters, and
streamf low for the rising and falling limbs of storm
hydrographs, Trap Bay Creek, Chichagof Island, Alaska,
during the Fall, 1980, storm season
102
LIST OF TA3LES - continued
Table
12
13
Page
Relationships between bedload transport, coarse particulate organic matter, two particle diameters, and
streamf low for individual storm events which occurred
during the Fall, 1980, storm season at Trap Bay Creek,
Chichagof Island, Alaska
105
Bedload (BLD) and coarse particulate organic matter
(CPOM) relationships for Flynn Creek, Oregon, and Trap
Bay Creek, Chichagof Island, Alaska
115
APPENDIX TABLES
Table
14
16-17
18-23
Page
Morphometric Characteristics of Trap Bay Creek, ChiChagof Island, Alaska...
...... 144
Relationships Between Bedload Transport, Coarse Particulate Organic Matter Transport, Two Particle Diameters, and Streamf low for Individual Storm Events
at the Upper and Lower Riffle, Trap Bay Creek, Chichagof Island, Alaska, During the Fall, 1980, Storm
Season
128-31
Relationships Between Bedload Transport, Coarse Particulate Organic Matter Transport, Two Particle Di.ameters, and Streamf low for Rising and Falling Limbs
of Storm Hydrographs for Individual Storm Events at
the Trap Bay Creek, Chichagof Island, Alaska, During
the Fall, 1980, Storm Season
132-43
Bedload Sediment Transport and Channel
Morphology of a Southeast Alaskan Stream
I.
INTRODUCTION
edload sediment transport in streams is a natural process
that removes the relatively coarse-sized products of erosion from the
site of weathering and moves them through the fluvial system.
bed-
load particles undergo further attrition and weathering during transport and may be altered so that they enter the suspended or dissolved
load portion of the total load of the stream (Swanson et al. in press).
The geology and geomorphology of a watershed and the climate, as it
deteriines the amount and form of precipitation characteristic of an
area, are the primary factors influencing the size distribution of
particles and the total sediment load.
In theory, the physical sta-
bility of a stream system is maintained by its characteristic fluvial
sediment load (Committee on Erosion and Sedimentation, 1977).
Iii-
creases or decreases in sediment load can initiate adjustments in
channel form by upsetting the dynamic equilibrium that exists under
the natural sediment regime, thus altering the physical and biological characteristics of the system (Heede, 1975; Park, 1976).
Many land-use activities that result in accelerated sedimentation
are considered to act as sources of non-point pollution.
Sedimenta-
tion in small streams represents an important non-point water quality
problem in southeast Alaska (Beschta, 1979).
Forest management activities are only one of the many that may
lead to accelerated erosion and alter the sediment regime of a stream
system.
In southeast Alaska, fish and timber are the two most
2
important resources at the present time (Meehan, 1974).
The fish-
eries resource is also the one most likely to be adversely affected
by any increased sedimentation arising from timber harvest.
Conflicts
between timber harvesting and fisheries are an important problem for
public land administrators (Beschta, 1979).
Research on sediment transport has largely bees limited to the
measurement and analysis of:suspended sediment, i.e., inorganic material transported by the turbulence of the stream and maintained in
the water column
This situation may reflect the relative ease of
design and use, as well as the relatively high efficiency of samplers
for the collection of this portion of the total sediment load.
High
levels of suspended sediment have been found to fill substrate interstices, and reduce cover and habitat for algae, aquatic insects, and
small fish (Nuttal, 1972; Brusven, 1980).
Fish ad other aquatic pop-
ulations may also be reduced or altered once suspended sediment is
deposited.
Deposition can reduce or block intergravel flow velocities
which are necessary to maintain a sufficient supply of dissolved oxygen for the respiration of fish eggs and invertebrates, and for the
removal of the waste products of metabolism of these organisms (Rynes,
1970).
Bedload sediment is defined as that portion of total sediment load
that moves by sliding, rolling, or saltating on or near the streabed,
moving at velocities that are less than that of the adjacent flow
(Harris and William, 1971).
This somewhat arbitrary definition makes
it difficult to separate the suspended and bedload portions of total
load.
Rocks and gravel would fall into the bedload category, while
3
sand-size particles could be transported in either mode, depending on
flow conditions and particle density.
Most field research on bedload
transport has been conducted within the last forty years, yet little
is Iaiown about its physical and biological relationships with the
stream ecosystem.
Land-use activities may have an effect on either
the suspended or the bedload regime.
A better understanding of bed-
load transport as it occurs under natural conditions is needed in
order to comprehend the interactions between land use, sedimentation,
and stream ecosystems.
Bedload studies in the field have generally been hampered by
a lack of adequate sampling devices.
The International Organization
for Staxidardization (ISO) has not yet established standards for bed-
load samplers or for sampling methods.
ISO has issued guidelines for
bed material sampling, but the spacing or frequency of bedload sampling
has not been specified.
Very little is known about the factors that
control the transverse variations of bed particle sizes at a given
cross section or how these distributions vary with time and flow conditions (Nordin, 1981).
The major problems involved in sampling bed-
load are: (1) There is no sampler that does not alter discharge ar
the streambed, or that is not selective in sampling certain particle
sizes; (2) obtaining reasonable average values requires the collection
of many samples because of the great temporal and spatial variations
in bedload transport.
A modified, hand-held, Helley-Smith pressure-dif-
ferential bedload sampler was used in this study to intensively sample
bedload transport during the high flow events of the Fall of 1980 in
a third-order stream in southeast Alaska.
Descriptions of this sampler
4
and its use are available in the literature (Drufel e.t al., 1976;
Eet, 1979, 1981; Helley and Smith, 1971; Johnsone.tal., 1977;
Beschta, 1981), and are discussed in the
Literature Reviews' section
of this paper.
Background and Objectives of Study
This study was sponsored by the USDA-Forest Service, Forestry
Sciences Laboratory (FSL), Juneau, Alaska, and conducted in cooperation with the Alaska Department of Fish and Game - Sport Fish Division
(ADF&G) and the National Oceanic ad Atmospheric Association - Nation
al Marine Fisheries Service O14FS)
The Trap Bay watershed was the site of research and background
data collection by the aforementioned agencies in the areas of benthic
invertebrate, insect, and fish populations; water temperature and
quality; precipitation and hydrology; and hillslope stability.
Trap
Bay Creek, a medium-sized third-order stream, and its tributaries are
located on the northeastern end of Chichagof Island on the southeastem
side of Tenakee Inlet.
The area is part of the land designated
for timber harvest in the Alaska Lumber and Pulp Company 1981-86 Timber Sale.
Cutting units and the main haul road have been surveyed
and staked out; two small tributaries to Trap Bay Creek drain cutting
units.
Details of the sale are available from William P. Gee, Forest
Supervisor, Chatham Area, Tongass National Forest, P.O. Box 1980.
Sitka, Alaska 99835.
5
The Watershed is moderately productive for pink sa1on
1
varden char; a small number of coho salion also spawn here
and Dolly
Old growth
Sitka spruce and western hen1ock are the major conerical timber
species.
Timber production is moderataly good, but much of the water-
shed is composed of very steep slopes of low-lying muskeg.
Sediment transport research included suspended and bedload sediment sampling on Trap Bay Creek during storm evetits from September
through November, 1980.
Channel characteristics were ieasured in Au-
gust of 1980 and 1981.
Objectives of the research were:
To quantify sediment transport rates, particularly the bedload component,
To relate sediment transport rates to those major hyrologic
parameters which appeared to determine the mechanisms of
sediment transport,
To evaluate how bedload transport influences channel
inor-
pho logy.
This study was an extension of a continuing research program on
sediment transport processes in mountain streams that is being conducted by the Forest Engineering Department, Oregon State University,
Corvallis, Oregon.
1See Appendix 1 for a list of the scientific names of flora and
fauna mentioned in this paper.
6
II.
LITERATURE REVIEW
Forest Management Activities and Fluvial Transport Processes
Physical and chemical erosion processes are continuously at
work at the aarths surface.
The products of these processes are
moved dowrislope by gravity, precipitation, wind, and biological ac-
tivities, where they become available for transport by streams
(Swanson, 1980).
Once fragmental materials enter the stream system,
bedload and suspended transport processes move particulate matter
through the channels.
Forest vegetation strongly influences nearly
afl. elements of the soil/sediment routing system on slopes and in
small streams (Swanson et a].., in press).
The amount of sediment that is eroded, transported, o
depsited
in a stream is a function of many interrelated variables, including
climate, soils, topography, vegetation, and land use.
In an undis-
turbed steep forested watershed, solution transport, the transport of
dissolved minerals, is the only perpetual transport mechanism0
Where
precipitation exceeds approximately 65 cm per year9 solution trans-
port can exceed particulate transport in terms of the volume of material exported per unit area (Clayton, 1981).
In contrast to the
generally perpetual transport of dissolved materials, periods of sediment movement range from frequent, low magnitude suspended transfer
to infrequent, high magnitude debris torrent events (Swanson et a10
in press).
The capacity of a stream to transport the sediment sup-
plied to it is a function of the hydraulic properties of the channel
(O'Leary, 1980).
7
Surface erosion (splash, sheet, rill, and gully erosion) is a
result of raindrop impact, thin film flow, or concentrated surface
runoff over the watershed (Satterlund, 1972).
It is rarely a problem
on undisturbed vegetated watersheds where the vegetation acts as a
buffer, absorbing energy from raindrops, and where the incorporation
of organic matter into the soil helps to improve the soil structure
and increase infiltration capacities.
The highly permeable soils
typical of southeast Alaska ensure that drainage is primarily by subsurface flow with little or no surface flow outside of established
channels (Swanston, 1974).
Dense vegetation and the thick orgatiic
layer present under the old growth forest are particularly effective
in protecting the soil from surface erosion.
The dominant geologic processes that transport the products of
weathering from the hillslope to the stream channel under forested
conditions in the steep terrain of the Northwest are soil mass movements (Swanston, 1974).
The entire soil mantle may be subject to the
set of processes termed ttcreep," which includes rheological deformation
and root throw (Swanson, 1980).
Debris avalanches and debris flows
occur on oversteepened slopes as a result of surface loading, increased soil water levels, removal of mechanical support, or a combination
of all three (Swanston and Swanson, 1976).
These types of mass move-
ment are the most frequently occurring types of mass failure in southeast Alaska and involve the rapid downslope movement of soil, rock,
and forest litter with relatively high water content (Swanston, 1974).
Soil creep may contribute to the susceptibility of soil to sliding in
critical areas (Swanston, 1974).
8
Debris flows and avalanches are usually in.itiated at the heads of
or within, shallow linear depressions on the valley side slopes (Swanston, 1974)
These hollows serve to concentrate soil seepage and de-
velop into surface drainages during major storms
In some instances,
debris avalanches may deposit their load at the base of a slope where
sediments can be supplied to the stream in small increments over long
periods of time.
Debris can also be carried directly into the stream,
producing a temporarily heavy sediment load or, under certain conditions, initiating a debris torrent which can scour the channel and
result in a large debris darn (Swanstot and Swanson, 1976).
In either
case, the sediment regime of the stream will be disrupted arid sedi-
ments may be deposited on and within spawining gravels where they can
disrupt the flow of oxygen to fish eggs and alevins and block the
emergence of fry (Chapman, 1961; Phillips, 1971).
Debris avalanches generally occur in relatively shallow, cohesion-
less soils in which the angle of internal friction, friction along the
sliding surface, and slope gradient control internal soil strength and
gravitational stress.
Soil saturation and active pore-water pressure
developient during major storms can substantially reduce soil strength
and decrease the critical angle of stability of the slope (Swanston,
1974).
Timber harvest on oversteepened slopes can affect stability in
two major ways.
First, tree removal increases the amount of water
stored in the soil through decreased evapotranspirational withdrawals.
This increases the length of time that the soil is fully saturated by
reducing the amount of water necessary to recharge the soil water
deficit (Harr, 1976).
Soil moisture levels may significantly affect
9
seasonal rates of creep and slp-earthf low movement.
Secondly, the
anchoring of the soil mass to the parent material by tree roots is a
major factor affecting the stability of oversteepened slopes.
Maxi-
mum decreases in shear strength of anchoring roots occurs three to
five years after cutting, and is probably the time when slopes are
most susceptible to mass failure during major storm events (Swanston
and Walkotten, 1974).
Sediment may also arise from the streai channel itself.
The
force of flowing water erodes the stream banks and bed, and this material may then be entrained, transported, or deposited by the streaiSediment is stored in floodplains, alluvial fans, point bars,
f low.
and in deposits associated with large organic debris (Swanson and Lienkaexnper, 1978).
Bed material transport reflects channel stability and
determines gravel bed composition (Milhous and K1ingean, 1973; Beschta
and Jackson, 1979).
The removal of large organic debris from streams,
either as part of the harvest or as part of a stream-cleaning operation, can result in the release of the equivalent of up to 15 years
sediment previously stored behind the debris acci.ulations (Megahan,
1975).
A study conducted in western Oregon showed that road building and
landing construction associated with timber harvesting can increase the
occurrence of soil movement by 20 to 350 times its rate of occurrence
in an undisturbed forest (Swanston and Swanson, 1976).
Clearcutting
can cause an increase of from 2 to 20 times over natural rates of
occurrence.
Increases in the sediment supply to a stream as a result
of mass failures can cause changes in the diversity and productivity of
10
insect life (Brusven, 1980), changes in the suitability of gravels for
fisheries production (Chapman, 1961; Hall and Krygier, 1967), and over
the long term, increased sediment production may reflect a loss in the
productivity of forest soils (Curry, 1973).
Sediment in Streams
The understanding of sediment processes is at a qualitative level
(Coittee on Erosion and Sedimentation, 1977).
Suspended sediment
has been better quantified than has the bedload fraction of total load;
suspended sediment measurement teachiques are well docuniented (Vanoni,
1975).
However, both suspended and bedload transport need to be meas
ured at the same time in order to adequately characterize the sediment
regime of a stream.
The bedload fraction may be small relative to the
suspended load fraction, but its movement has a greater effect on channel characteristics.
A lack of reliable sampling methods for bedload
has been a major limitation in field studies; data analysis and comparison are further complicated by variations in research techniques
(Edwards, 1979).
Sediment transport begins when the force of streamf low acting on
channel materials exceeds the critical condition for motion.
This
concept of a critical tractive force, t, was defined by Dii Boys in
1879 as follows:
t = Y RS
where,
= specific weight of water
R = hydraulic radius
S = energy gradient.
(1)
11
Material that is entrained in the flow travels either as suspended
or bedload sediment.
Particles are suspended if the ratio of fall vel-
ocity, W, to shear velocity,
,
ed as bedload if the ratio W/
is less. than 0.8.
They are transport-
is greater than 0.8 and the Shields
criterion for initial motion is met, which for fully developed flows
is t/Yd
0.06 (Nordin, 1981).
pgDS
where,
(2)
p = fluid density
g
acceleration due to gravity
D
flow depth
S
slope of energy gradient
d
particle diameter
(p2-p)g
p2
sediment density
= (t/p)½
In most streams, suspended load comprises a larger fraction of
the total load then bedload.
Most mountain streams are supply-limited
in their suspended load, which is dependent on the amount of fines
(silts, clays, very fine sands) present in or transported to the
stre.
The bedlo.ad supply, in contrast, is generally greater than
the stream can transport.
Thus, bedload transport occurs only during
periods of high flow, and then only over relatively short distances.
Table 1 siimmrized the findings of several researchers on the
percent of total load transported as bedload.
It can be seen that for
several streams in the Pacific Northwest, bedload comprises only about
one to 25 percent of the total load transported at peak flows.
This
fraction may be somewhat larger for mountain streams and tends to
12
TABLE 1.
Percent of Total Load Transported as Bedload During
Peak Flows for Several Streams in the Pacific Northwest
Region
% bedload at peak
discharge
Source
Idaho Batholith
(Upper Salmon River)
Eett, 1975
N. Coastal California
(Van Duzen River)
Kelsey, 1977
Oregon Coast Range
(Oak Creek)
Klingeman &
Milhous, 1970
Oregon Coast Range
(Flynn Creek)
Edwards, 1979
13
increase as discharge increases (Klingeman and Milhous, 1970).
Bedload transport is influenced by channel slope and roughness,
particle size, and stream velocity and turbulence (Simons and Senturk,
1976).
Changes in any one or several of these factors alter the trans-
port capacity of the stream, and may result in aggradation or degradation of the channel.
Furthermore, changes in the sediment supply due
to increased erosion in the upper reaches of the watershed can result
in general deposition of the sediment and aggradation alotig the lower
channels (Leopoldetal., 1964).
Bedload Measurement and Prediction Techniques
Direct sampling atid measurement methods were used for estimating
bedload tranpsort up until the early 1940's, when Einstein introduced
his empirical bedload equations (O'Leary, 1980).
Development of n-
erous theoretical atid empirical bedload equations followed.
Attempts
were made to construct bedload transport equations that would predict
the maximum tratisport capacity of a stream for a given set of hydraulic cotiditions and sediment characteristics (Graf, 1971; Vanoni, 1975).
A number of investigators have studied the applicability of these
equations (Klingeman, 1971; Haddock, 1978).
The models studied were
found to inadequately represent high-energy mountain streams due to
the supply limitations which arise from the flushing, deposition, and
armoring of the streambed (Milhous and .Klingeman, 1973; Haddock, 1978).
Most empirical and theoretical equations have been developed from flume
studies or for relatively constant conditions which are not character-
istic of natural streams; thus they do not adequately represent the
14
the conditions under which natural sedinent transport normally occurs.
These problems have prompted a return to direct sampling techniques.
Of the bedload sampling devices developed prior to the 1940's,
there are two categories of samplers which come close to meeting the
three criteria for a
ideal bedload sampler listed in the 1940 U.s.
Federal Inter-Agency River Basin Report:
The sampler must sample a definite portion of the moving
water and solids.
All the solids moving in the sampled portion must be collected.
The sampler must have a secure contact with the bed surface and not disrupt upstream flow nor obstruct the entrance of particles.
These two samplers are the pressure differential and the vortex
samplers, although problems have been observed with each type of
sampler.
The vortex sampler had been found to underestimate the
actual bedload transport rate.
particle sizes of less than 4.75
It is especially inefficient for
in diameter and becomes more in-
efficient as discharge increases (Iüingeman, 1971; Hayward and 5utherland, 1974).
The pressure differential sampler
ay either over-or
underestimate bedload tranpsort depending on which particular device
is used (Hubell, 1964; Helley and Smith, 1971).
An additional prob-
lei with the pressure differential sampler is that its use requires
moving the device, which can further disrupt flow.
Good contact with
the streambed may not always be possible and local scour can occur
at the point of placement (Helley and Smith, 1971).
15
Use of a vortex sampler requires that the device be installed
in a uniform cross section of the stream.
Installation costs are
high and the vortex sampler is not transportable from place to place.
Pew researchers have used vortex samplers for these reasons (OtLeary,
1980).
There have been .f our studies published so far in which the
researcher(s) used a vortex spler:
Klingeman and Milhous (1970),
Hayward and Sutherland (1974), O'Leary (1980), and Edwards (1979).
The latter two included comparisons of vortex and Helley-Smith pressure differential sampler efficiencies.
The lack of a uniform, stable cross-section for installation,
relatively high installation costs, and inaccessability of the research area made the use of a vortex sampler infeasible in this study.
The Helley-Smith pressure differential sampler has been used in a
number of studies of natural bedload transport and a slightly modified version of it was chosen for use in Trap Bay Creek.
Pressure differential samplers theoretically equalize the entrance
velocity of the sampled portion of streamflow and that of the surrounding stream through a pressure drop created by the divergence of the
exit walls of the sampler.
As the velocity decreases at the downstream
end of the sampler, the sediment in transport is deposited.
The Dutch,
or Arriheim, sampler was the most widely used model prior to the develop-
ment of th
Helley-Smith.
The Helley-Smith sampler is a modified ver-
sion of the Arriheim spler and was designed for use by the U.S. Geologic Survey.
It has a 7.6 cm
square aperture and was designed to be
used in flow velocities of less than or equal to three meters per second
(m.s1)
(Emmett, 1981).
Efficiency tests of the Arnheim and Helley-Smith
16
overestimates by about 50% (Hubbell, 1964; Helley and Smith, 1971).
The variability in sampler efficiency was originally thought to
be a function of the bed material used in testing, a 1.15 = diameter
sand, and the natural variability of bedload in time and spaca.
In
his 1971 report to Helley and Smith, Jobson1 stated that he had found
that the sampler tended to slide upstream a small distance while being
raised from the bed, and would thus tend to scoop up additional sand.
He concluded that the scooping tendency might be reducad by using
larger gravel in efficiency tests, and the sampler would then give a
better estimate of transport rates.
More recent studies have shown
that the overestimation of transport rates is a result of increased
velocities at the orifice, while underestimation is a result of the
clogging of the mesh collection bag on the sampler.
The hydraulic
efficiency of the sampler (the ratio of the mean velocity of water
through the sampler to the mean velocity of water had the sampler
not been there) was found to be about 1.54 (Emmett, 1981).
Emmett
(1981) also states that the sampling bag can be 40¼ filled with sedi-
ment which has a particle size larger than the mesh size without losing
efficiency.
However, smaller particles tend to plug the bag and are
subsequently lost from the sample.
The problem of the clogging of the sampler bag was the subject
of a 1981 study by Beschta.
Both Beschta (1981) and Johnson et al.
(1977) observed that fine sands and particulate organic matter tend to
clog the mesh of the sampling bag.
11N:
Helley and Smith, 1971.
This reduces the effective flow
17
through area of the sampler,creating a back pressure at the orifice and
reducing efficiency.
Beschta (1981) showed that trapping efficiency
of a standard collection bag (surface area 1950 cm2, 0.2 = mesh size)
exposed to a 0.50 = (average particle diameter) sand mixture in a
flume was a function of the length of time of the sampling period.
The use of a larger collection bag (6000 cm2, 0.2 = mesh) resulted
in the efficiency remaining constant for sampling periods of up to
eight minutes.
Sampler efficiency would probably decrease for samp-
ling periods longer than this, or if large amounts of particulate organic matter were in transport.
Studies of bedload transport have been made using the HelleySmith sampler in the field as well.
Molnau et al. (1975) used it in
the Knapp and Cape Horn Creeks of the Central Idaho Batholith.
streams have beds consisting of sand and fine gravel.
These
For Cape Horn
Creek, the researchers found that bedload transport rates increased
on the rising limb of the hydrograph, dropped sharply prior to the
peak, and increased again on the falling limb.
Molnau et al. (1975)
explained this rising limb shift based on the tractive force theory1.
The tractive force necessary to initiate bedload tranpsort was exceeded early in the snowmelt season,causing the increased transport
on the rising limb.
The drop in bedload transport was hypothesized to
indicate that the stream had cleansed itself of all tttransportablet
sediment, that sediment deposited on the armor layer over the previous
year or time since last critical discharge.
Thus, although the
1See pages 12 and 13 for explanation of tractive force theory.
18
tractive force increased further at the peak of the hydrograph, no in.
crease in bedload transport could take place because of a lack of
transportable sediment.
In Knapp Creek, transport rates also increas-
ed on the rising limb of the spring melt hydrograph, but a second
sharp increase occurred just prior to the peak of the hydrograph.
Molnau et al. (1975) hypothesized that discharge had increased to the
point where the tractive force was great enough to dislodge the armor
layer, releasing sediment that had been trapped below it.
According to Beschta (personal coimnunication), Molnau et al.
(1975) based their hypothesis on a relatively few samples.
Much of
what they saw may actually have been sampling variability.
In the fol
lowing study by Etett (1975), a ten-fold variation in peak bedload
transport rates was cotnmon.
ett (1975) used a Helley-Smith sampler to collect bedload data
for three streams of the upper Salmon River basin of Idaho.
Although
there were too few samples taken to detect differences in transport
rates prior to or after the peak discharge, bedload transport was
found to increase with increasing discharge0
Measurements were taken
in Slate Creek to obtain average minimum transport rates at the
cross-sectional location where the maximum transport rates were observed.
Over a three day period, the average rate of transport was
approximately the same for the same discharge.
Individual transport
rates comprising these averages showed about a ten-fold variation, however0
These variations were related to some hydraulic variables, but
the relationships were not consistent.
Samples collected from the Snake and Clearwater Rivers in Idaho
19
were used to estimate bedload transport rates in kilograms per meter
width of channel per second (kg.ms)(iiett, 1976).
This data
was plotted against stream power, as defined by Bagnold i:1977), also
in
kg.m1s1.
This graph showed that one relationship was applicable
for low values of stream power when coarse particles are not moving and
fine particles are limited, and that another relationship was applicable at higher values of stre
are moving.
power when almost all bed materials
At intermediate values of stream power, there is a break
in the relationship. Eett felt that the lack of intermediate sized
gravels in the bed material of these two rivers (binodal size distribution) led to this; the bed was either artored or moving.
The results of a field calibration conducted on the East Fork
River, Wyoming, in 1979, were presented by Eett (1981).
Helley-
Smith samples were compared with results obtained using a conveyorbelt sampler which was assumed to be 100% efficient.
of large particles, 8 to 16
Transport rates
in diameter and 16 to 32
n in diameter,
were too minimal to allow calibration for particle size ranges greater
then 16
.
Direct comparisons with results obtained using the con-
veyor belt could be made for particle sizes ranging from 0.5 to 16
m.
Eet felt that the' Helley-Smith sampler was 100% efficient for the
0.25 to 0.50
less than 0.25
than 16
particle size class, but that data on particle sizes
should not be used and that data on particles larger
n should be treated with caution.
Using data from the East Fork River, Eett (1981) quantified
bedload transport in terms of the particle size classes as a percent of
total load, and the rates of change of a given size class as the total
-
20
bedload transport rate increased (Table 2).
A least squares linear
regression of the log transformed data was used to develop a power
equation relating the bedload transport rate in each particle size
class (Y) and the total bedload transport rate (X):
(3)
"b", the slope of the regression line, is the rate of change in per-
centage of total bedload in that particle size class.
The Helley-Smith sampler does pick up some suspended sediment,
the absolute quantities of which depend upon the sizes of sediment
in transport and the hydraulic conditions of flow (Eett, 1981).
However, the significance of material that can be transported as
suspended sediment (less than 0.50
bedload transport rate increases.
in diameter) decreases as the
The rate of change values (exponent
b) in Table 2 for suspended size particles are less than imity, indicating that the percentage of sediment in those size classes decreases
as total bedload transport rate increases.
The significance of transport of particles in size classes rang-
ing from 05O to 8.0
was found to be greatest and to increase as
bedload transport increased (Table 2).
class was 0.50 to 1.0
.
The dominant particle size
The greatest rates of change occur from the
2.0 to 4.0 = size class, followed by the 1.0 to 2.0 = and 4.0 to
80
size classes.
At high bedload transport rates, the rate of
change values combine with the mean percentage values such that the
1.0 to 20 and 2.0 to 4.0
centage of total bedload0
size classes comprise the greatest perThe median particle values or rate of
change for the two coarsest particle size categories are not comparable
21
TABLE 2.
Mean Percentage of Total Bedload in Each Particle Size Class
(Y/X) and Rate of Change in Percentage (B) as Bedload Transport Rate Changes1'2
Particle size
class, =
Mean percentage of
total bedload in
particle size class
(Y/X in %)
Rate of change in
percentage of total
bedload in particle
size class (B)
0.06 - 0.12
0.35
0.727
0.12 - 0.25
3.24
0599
0.25 - 0.50
22.80
0.698
0.50 - 1.00
26.84
1.050
1.00 - 2.00
20.07
1.213
2.00 - 4.00
10.61
1.344
4.00 - 8.00
3.45
1.193
8.00 - 16.00
0.89
0.867
16.00 -32.00
0.65
0.367
1Data adapted from E=ett (1981).
x
= bedload transport rate in a given particle size class.
total bedload transport rate in a given particle size class.
22
to those for the smaller size categories because the largest particles
move only at high transport rates.
Thus, many of the low transport
rims could not be included in the analysis for these size particles
(log transformed regressions cannot include zero values) (ett
1981).
Eiett (1981) concluded that the Helley-Smith sampler should not
be used for measuring bedload transport rates for sediment of particle
sizes which also are transported as suspended sediment; nor where the
bag may become clogged with particles about equal in size to the size
of the mesh, or with organic debris0
Sampling efficiency probably al-'
so decreases as particle size appruaches nozzle dimensions.
Further,
the sampler should not be used when irregularities in the stream bed
preclude a reasonable fit between the sampler boto and the bed0
Edwards (1979) conducted sediment transport research and used
both a vortex sampler and a hand-held Helley-Smith sampler.
was conducted on Flynn Creek, which drains a 2
shed in the Coast Range of western Oregon
The study
undisturbed water-
The Helley-Smith sampler
was found to be more efficient in this stream, where much of the bed-
load consists.of sand-sizematerial.
The Helley-Smith sampler is also
compatible with standard suspended sediment sampling techniques, and
it can be modified to account for local sampling conditions.
This is
advantageous in Flynn Creek because of the variation in transport
rates and types of material in transport0
Edwards found such differ-
ences existed among three sites selected for bedload sampling, mdicating that limited sampling at a single site
acterize bedload transport in this area.
ay not adequately char-
23
Edwards (1979) also found that bedload transport occurred in
relatively discrete pulses.
coincide with peak runoff.
Maximi.mi bedload discharge did not always
Although streamfiow was found to be the
principal variable controlling sediment transport, results indicated
that supply limitations exist.
Suspended sediment represents the
greater portion of total load in Flynn Creek, yielding from four to ten
times more sediment than was yielded as bedload during a 24-hour peak
flow period.
Both total suspended solids (TSS) and coarse particulate
organic matter (CPOM) peaked early in the storm; however, pulses of
CPOM discharge occurred during the recession and appeared to be related
to streambed disturbances.
Edwards concluded that channel morphology
and in-stream obstructions may cause significant spatial and temporal
variations in sediment transport.
Bedload movement is important in
regulating bed composition and particulate yields.
Another study by O'Leary (1980) in Flynn Creek used a vortex
sampler as the primary device for sampling bedload tranpsort during
storms.
A hand-held Helley-Smith sampler was used to obtain samples
composed of several cross-sectional subsamples, which provided supplemental information about bedload transport, and provided "test samples'
which were used to measure equipment efficiency or variations in transport across the channel.
Samples collected with the vortex sampler in-
dicated that bedload discharge increased with increasing streaniflow.
Bedload transport rates also increased as the storm season progressed,
even though streaniflow was less than that of previous storms.
O'Leary
(1980) concluded that this may have been the result of exceeding some
critical discharge necessary to initiate bedload transport.
Regression
24
analysis of streamflow and bedload discharge showed that significant
(p - 0.01) differences existed between relationships developed for
individual storms.
Helley-Smith samples obtained by O'Leary (1980) indicated bedload discharge rates greater than those indicated by the vortex
sampler, except for one storm.
This was again attributed to the
greater efficiency of the Helley-Smith for material in the sand-size
particle range.
Intensive sampling with the Helley-Smith sampler
showed that bedload transport rates fluctuated rapidly.
O'Leary
(1980) also found that when large amounts of organic material were in
transport, bedload transport rates as estimated by the Helley-Smith
samples tended to be low.
However, there was no statistically sig-
nificant regression relationship between percent organic matter and
the bedload transport rate.
Another method that has been used to estimate bedload transport
has been to measure changes in the channel morphology.
This has been
done by measuring cross-sectional profiles prior to and following the
storm season or, in some cases, following individual storm events.
Cross-sectional profiles and longitudinal morphometric surveys have
been combined to show net channel changes over time (Leopold et al.,
1964; Dunne and Leopold, 1978; Edwards, 1980).
Various morphological characteristics have been used in attempts
to predict sediment discharge (Rosgen, 1978; Marston, 1978; Dunne and
Leopold, 1978).
Narston (1978) used data on the morphometric charac-
teristics of several streams in western Oregon to develop regression
equations to predict streataflow and sediment yield.
Rosgen (1978)
25
shows how sediment rating curves can be used to predict changes in
sediment yield following silvicultural activities.
Both of these
methods have limited usefulness, however, because of the large amount
of data which must be collected and because the relationships tend
to be site specific.
Tracers, in the form of marked rocks, have been used to relate
discharge to the nunber and size of bedload particles moved during
high flows (Leopold and Fnett, 1981).
Rocks representing various
size particle classes were collected from a streambed, marked, and
replaced in a line along a riffle.
The authors found that even when
the computed shear stress was several times larger than the minimal
value for motion as indicated by initial motion stress values derived
from the Shields curve, only a small proportion of available rocks
would actually be set in motion.
Therefore, to move all rocks of a
given size, it is necessary to have repreated flows which produce a
competent shear stress.
Leopold and Fett (1981) stated, "If a gra-
vel riffle is an expression of a kinematic wave as suggested by Langbein and Leopold, complete replacement of rocks in a zone of concentration requires a series of flow events of sufficient energy to move
those rocks."
Thus, although a riffle may represent the crest of a
wave of bedload transport, bedload movement itself is a function of
more than the shear stress resulting from streamf low.
Color-coded marbles were used by O'Leary (1980) to determine bed
scour and fill, and average transport distances of bed material.
Al-
though the marbles were resistant to abrasion, so that the color
would not come off, and they approximated bedload particles in size
and density, difficulties arose in recovering them following the storm
26
season.
Three sizes of marbles were used in the study; percent re-
covery was significantly higher for large and medium-sized marbles
than for small marbles.
The method did show where bedload movement
and channel changes occurred
but results obtained for transport in
terms of distance moved and number of marbles moved may have been
biased towards the larger marble sizes (O'Leary, 1980).
Problas associated with using "tracers" to determine transport
distances tend to be related to recovery problems.
subject to abrasion and lose their markings.
Painted rocks are
Even if this is not a
problem, 100% recovery of tracers is often difficult.
The use of
materials to mark tracer particles may be one way to overcome these
problems.
None of the methods available to evaluate bedload transport in
streams is entirely satisfactory.
Theoretical and empirical equations
do not adequately characterize conditions occurring in natural streams.
Direct measurement techniques utilize samplers that alter streamflow
conditions, which leads to an overor underestimation of bedload
transport.
Direct sampling techniques and measurements of chaxiges in
charmel morphology are very labor intensive; large amounts of data are
required to yield reliable results.
However, the use of these various
techniques will hopefully lead to a better understanding of sediment
transport under natural conditions, and aid in the development of more
suitable methods to evaluate changes in the sediment regime in streams
27
III.
WATERSHED DESCRIPTION
The Trap Bay watershed is located on the southern side of the
Tenakee Inlet, on the northeastern side of Chichagof Island, between
57°44' and 57°45' north latitude and 135°00' and 135°02' west longitude (Figure 1).
It is approximately 60 aerial miles SSW of Juneau,
Alaska, and is 13.5
14ii2
2)
in area.
The climate is typical of the Alaska Panhandle and is a cool,
moist maritime climate characterized by cloudy skies and little daily
temperature variation.
Cloudy skies occur on the average of 275 days
per year, 43 are clear, and the remainder are partly cloudy (Harris et
al., 1974).
The relatively wide range in daily hours of sunshine dur-
ing the year apparently reduces daily temperature fluctuations.
During
the ser, there is only a brief period of nightime cooling, while
du.ring the winter, the low angle of the sun and reduced hours of sun-
shine result in little surface heating (Harris etal., 1974).
other major moderating influence on temperature is the sea.
of the Inland Passage are warmed by the Alaska current.
The
Waters
Sea temper-
atures range from 12.8°C (55°F) in the ser to about 5.6°C (42°F)
in the winter (Harris etal., 1974).
The watershed receives about 3410 txmi (95 in) of precipitation per
year, with nearly 40% occurring in October, and only 1% occurring in
April, May, and June (Harris etal., 1974).
Most precipitation occurs
as steady, light to moderate rain, although there may be appreciable
snowfall at higher elevations during the fall and winter.
Maximum precipitation is usually associated with prominent low
pressure systems, called Aleutian Lows, which develop in, or cross,
28
:yp, ":
'\
,
-.)
:.
.
2k-.
?'
-
.-
j
r:.-.
-
7
ç-
;;i_-
-
'ç
-
ø4/
?w1
/ACIr'
'--
.LT
TAU-
-cz,
'
Hi-(.-
-'.,'-.C
-
l,u'.
-
PIEAS4N\
C.
I
S.'
_e.f,
.,,
SSLANO
.
PMs Lo..nj.
alsic.,
'Joe.
5',ouczè
-
5'
C,
'.
'-
.
C,,f Poh1
p.J.
-..' ,
-
i:'--. DoyCov
--,
':-.N' Grc1.i iQd -'
*GSV's5'-
.-T
-
?
\'
--...5/
ilendc
I
aue.
-
'C'
.
.
'
-
-S
.
-4
L
\\
.
./'
'5_.
F..L'*lV
.............5'
-:
uIL4-\ -. .:..\
011W
S..
::
-i_.;
L
ADMIRALTY
n.nfl,u1j0
-
..ISLAND.
ak, \
POffi
4:
-
- .'-, 4_
.q,
C..Du',
VL-::hI
- ----------------
34SLAND-
-
-
'5
Figure 1.
-
l..
_i
\
;._
.5,.
.:-Teoi.
.'
ovasou'
r
-'
W.IIO
:
--f-
(. -5
l.jon
- ..j
General map of Chicagof Island showing location of
Trap Bay Watershed
0iflf Nooa.on
29
the Gulf of Alaska (Harris et al., 1974).
These systems follow a storm
track along the Aleutian Island chain, the Alaska Penninsula, and the
Gulf of Alaska.
These areas are exposed to most of the storms cross-
ing the north Pacific.
Moist air masses moving in from the sea are
lifted by the Coast Mountains, which interrupt surface air circulation,
resulting in the large amounts of rainfall that soak the Panhandle.
Trap Bay Watershed is a glacial c±rque valley bounded by serrate
ridges and a horn peak at the southern end.
Elevation ranges from
sea level to a maximi.
Detailed mapping and
of 1320 m (3870 ft).
interpretation of the geologic history of southeast Alaska is still in
the initial stages.
In most glacial cirque valleys in the area, the
bedrock plays a lesser role in soils development because it is often
overlain by compact glacial till up to 451 m (1500 ft) in elevation
(Harris etal., 1974).
Soil formation began following the retreat of
the last glaciers, which were associated
ith the Wisconsin advance;
most tnineral soils are derived from ablation till (Harris et al.,
1974).
Climate is an important factor in soil development.
fall, cool temperatures, a short growing season, an
High rain-
moderately low
soil temperatures all contribute to an accumulation of organic matter
on the surface and within the soil (Harris et al., 1974).
The organ-
ic mat on the surface, for example, may range from 15 to 25 cm in
thickness.
Sidel and Swanston (1982) described the soils of a northwest
facing slope on the eastern side of the watershed.
Mid to upslope
soils are 15 to 50 cm thick and are overlain by a wet, dense, organic
30
layer which is about 20 cm thick.
The Tolstoi soil series predominates
in well-drained slope positions and the St. Nicholas series occurs in
more poorly drained steep sites.
er and midslope reaches.
The Kupreanof series occurs in low-
These soils are moderately deep, well drain-
ed, and overlie weathered graywacke.
Soils common to s.milar areas of southeast Alaska contain approximately 1QZ organic matter and 12% iron oxides
both of which strongly
attract and hold water (Harris etal., 1974).
Thixotrophic properties
are coon in many southeast Alaskan soils as a result of the high organic and iron content in combination with high rainfall.
is a reversible gel-sd
Thixatrophy
transformation in which the soil structure
breaks down under stress.
Precipitation generally exceeds calculated evapotranspiration
throughout the year.
This, in combination with the many glacially-
caused depressions and extensive impermeable soil layers, has resulted in the formation of large areas of organic soils.
These are classi-
f led as Histosols and coonly called "muskegs" (Harris etaL, 1974)
Muskegs cover much of the lowland area of the watershed.
They are
composed of sedge or sphagntnn peat, and support sedge, sphagnum, erica-
ceous species, and stunted lodgepole pine and western hemlock.
Muskegs
may help to regulate streamf low (Harris etal., 1974).
All mature mineral soils under timber have strongly spodic (Pod.'
zol B) horizons (Harris et al., 1974)
Spodosols have developed under
the spruce forest which covers the slopes and better drained lowland
areas.
organic
The depth of tree rooting is largely confined to the thick
at and the upper 30.5 cm (12 in) of mineral soil where most
31
plant nutrients are concentrated.
Vegetation is rich arid abundant everywhere in the watershed except above treeline.
I observed Sitka spruce, western hemlock,
and scattered red-cedar and red alder in the forest.
The understory
consisted of blueberry, huckleberry, ferris and numerous vascular
plants.
Mosses and lichens covered every available surface.
Dense
thickets of salmonberry, ferns, skunk cabbage, and nettles alternated
with alder clones and an occasional hemlock along the streambanks and
in frequently inundated areas.
Devils club was also found along
streams, throughout the forest, and was particularly abundant in
clearings and on steeper slopes where the soil becomes thin and rocky.
A variety of grasses grew along the shorline, extending inland as far
as the high tide line.
Descriptions and geographic ranges of most
woody species of the region can be found in Alaska Trees and Shrubs
(Viereck and Little, 1972).
The watershed is characteristic of southeast Alaska with its
steep slopes; shallow, highly permeable soils; and high rainfall.
Slopes range from 5% in the valley to about 75% along the side of the
ridges.
The dense vegetation and thick organic layer effectively
protect the soil from surface erosion.
Soil drainage is primarily by
subsurface flow; high soil permeabilities ensure that little surface
flow occurs outside of established channels.
Oversteepened slopes,
those having a slope angle of greater than 30 degrees, are subject to
creep which may contribute to the susceptibility of soil to sliding
in critical areas (Swanston, 1974).
Also, the watershed is in a zone
of seismic activity, and is undergoing tectonic uplift following the
32
last glacial retreat (Swanston, personal coiiunication).
Ongoing re-
search which is being conducted by the Forestry Sciences Lab should
provide more inforxnation on the contribution of mass failure to the
sediment load of the drainage network, both under the virgin old-growth
timber and following timber harvest.
Stream Characteristics
Trap Bay Creek is fed both by a spring originating in a cave on
the eastern end of the watershed, and by a second-order stream draining the southern end of the watershed. Numerous small tributaries
flow into these two branches and into the main stream itself.
Several
of these tributaries flow through muskegs which contribute large
amounts of dissolved organic matter and result in discolored water.
Peak flows in the general region occur following the recharge of
the soil moisture deficit during the fall rains, and also during early
spring due to rain-on-snow events (Schmiege etal,, 1974).
Stream-
f low is at its lowest stages during the months of July and August when
evapotranspirational demands are highest and precipitation inputs are
relatively small. Lag time following precipitation varies with antecedent conditions; the hydrograph begins to rise approximately six
hours after the onset of precipitation. Rapid fluctuations in streamflow are probably reduced to some degree by the regulating influence
of the spring in the cave and by the muskegs.
In the upper reaches of Trap Bay Creek, channel form is primarily
influenced by large organic debris in the channel and streamside
tation. Pool-riffle sequences are generally a result of stable
vege-
-
33
accumulations of debris against fallen trees which create settling basins and tend to obscure any natural systemic poàl-riffle sequences.
Large trees and root wads frequently create protected backwater areas
that are important as habitat for fry during high flows (Swanson,
1980).
Downstream reaches are not as heavily influenced by the forest,
but large organic debris continues to play a significant role in channel form.
Streambanks become far less stable and are subject to fre-
quent sloughing during high flows where the protection afforded by
tree roots is no longer present.
The downstream reaches are also
affected by tides which vary from less than 0.3 m (1 ft) to more than
6.1 m (20 ft).
The effects of 6.1 m tides extend nearly 1280 m
(4200 ft) upstream and, when they occur in combination with high flows,
can result in the flotation of otherwise stable debris.
Beaver activity is coon along the lower 1890 m (6200 ft) of
channel.
Numerous piles of debris have been constructed which usually
divert flow.
A network of trails and slides also occurs along th.e
stream which may contribute to localized areas of bank instability.
During the months of August through November, and especially dur-
ing the major pink slamon run of late August, the streambed is disturbed by the spawning activities of the fish.
Pink salmon redd con-
struction redistributes large amounts of gravel and can change the
streambed profile drastically in heavily used areas.
Changes in the
suspended sediment load during these period may result from this spawning activity.
The longitudinal profile of Trap Bay Creek closely parallels the
34
valley gradient.
The lower 1372 m (4500 ft) of stream channel has a
relatively steady gradient of 0.25%.
Gradient decreases to about 0.18%
through a 168 m (550 ft) relatively straight section, and then begins
to increase.
This increase in gradient begins above the tidal in
fluence zone.
Gradient in a reach examined in this study
1591 m to
1409 m (5218 to 5279 ft) upstream of the stream mouth, increased from
0.41% in 1980 to 0.81% in 1981 (see discussion of channel morphology
in Results section).
The streambed is generally composed of small to medi
cobbles,
gravel and coarse sands with silt and fine sands increasing in abundance in depositional areas
low flows.
Gravel bars are nterous and obvious at
Sections of the streabed are armoured by cobbles that
range in size from one to several
in diameter.
Sand and gravel un-
derlie and fill the interstices between the cobbles.
In one short and
relatively straight reach of the stream (1340 to 1522 m from the stream
mouth) larger stones and boulders, ranging from 5 to more than 30 cm
in diameters compose part of the bed.
They increase in size and num-
ber in the upstream direction9 reaching a maximum in the vicinity of
a USFS stream gage, which is located 1521 m (4990 ft) from the stream
mouth.
The primary study site was located between 1590 and 1609 m from
the stream mouth, arid consisted of two riffles separated by a depositional area.
It was selected because it was the only pool-riffle se-
quence close enough to camp to enable manual transport of equipment and
materials, and it was only slightly affected by larger organic debris
There were several trees that had fallen across the stream along the
35
study reach, but all appeared to have been in the same position for
several seasons.
Only one of these trees, located at the upper riffle,
was actually on the streambed; this log was more than half buried in
gravel and there was gravel acci.ulating upstream from it, so it appeared to be stable.
bankfull.
The upper riffle was 13.7 m (45.1 ft) wide at
The lower riffle had a bankfull width of 12.6 m (41.4 ft).
The depositional area was approximately 18 m (60 ft) long and 10.7 m
(35.0 ft) wide.
A bridge was constructed across each of the riffles so that bedload samples could be obtained without disturbing the bed.
Some
riparian vegetation was cut down to acconodate construction.
Bed ma-
terials in the study section were typical of the stream, with an increase in the proportion of sand and silt size particles in the depositional area.
Stream banksaveraged 0.76 m (25 ft) high through the primary
study section, ranging from 0.17 m (0.5 ft) to 1.22 m (4.0 ft).
The
left1 bank appeared to be realtively stable due to the presence of the
roots of saitnonberry, alder, and hemlock.
Some undercutting was oc-
curring below the rooting depth of this vegetation.
Control Creek, a
-
tributary stream that entered the mainstream just below the foot of the
upper bridge, had been daned by beaver activity prior to 1979 (Hub-
bard, personal conunication).
It now joins the mainstream at two ad-
ditional places: at the foot of the lower bridge, and 30.5 m (100 ft)
downstream from the lower bridge.
The diversion of Control Creek has
1Left and right designate the side of channel relative to an observer facing in the downstream direction.
36
resulted in the year-round inundation of the lowland area adjacent to
the mainstream and the accelerated erosion of the right bank.
37
HETHODS
IV.
Precipitation and Streamflow
Two weighing precipitation gages (Weather-Measure Model No. P511P
Remote Recording Snow Gage) were installed on the watershed in the
Spring of 1980.
One was located in a natural clearing on a southwest
facing slope at about 150 m (490 ft) in elevation.
The other was io-
cated in a meadow, less than 400 m (1300 ft) from the stream, at an
elevation of about 15 m (150 ft).
Other studies have indicated that
there may be signficant differences between precipitation falling on
slopes and that received in valleys (Schmiege etal., 1974).
Mechani-
cal problems with the meadow gage made it impossible to determine if
this might be the case at Trap Bay.
A continuous record of rainfall
during the sampling period was available from the slope gage, except
for the rainfall of 24 September 1980, and was used to relate rainfall
duration and intensity to storm runoff.
A wate-level recording stream gage (Fischer-Porter Series 1540,
Model No. 35-D; accuracy ± ½ count) was installed at the head of a
chute, 1520 m (4990 ft) from the stream mouth, in July of 1980.
This
instrument recorded the water level at 15-minute intervals to the nearest 0.3 cm (0.01 ft) on a punch tape.
It was set to punch at the same
level as the stage indicated by a staff gage that was adjacent to it.
Two additional metal staff gages were installed below each bridge to
provide supplemental readings.
The float-counterweight fell off the
recording gage during the period from 11 October - 16 October, and during this time stage was determiaad from the staff gages.
38
A series of velocity measurements were made during September and
October, 1980, using a Teledyne-Gurley Direct-Reading Current Meter.
A regression equation was developed to relate stage to velocity for
each of the cross-sections below the bridges0
The stage-velocity re-
lationship was then combined with determinations of the cross-sectional areas occupied by water at a given stage and stage-discharge relationships were developed.
Regression relationships between stage at
the recording gage and the estimated discharge at each of the crosssections underneath the bridges allowed development of a stage-discharge curve for the entire stream.
Although there wera insufficient on-site data available to use
any of the coonly employed methods for evaluating the magnitude of
storm events, equations have been developed !or estimating mean monthly, mean annual, and peak flows of various return periods, based on
certain characteristics of the watershed (Water Resources Atlas for
Alaska, 1979).
These equations are included in Appendix 2.
Channel Morphology
A theodolite survey of the thaiweg was conducted during July and
August, 1980, from the mouth of the stream (assumed to be sea level at
station 0 at low ti4e mark on 22 July 1980) to 1646 m (6400 ft) upstream.
Thalweg elevation, bankfull width, and the distance of the
thaiweg from the right bank were measured at 15.2 m (50 ft) intervals.
Stakes were placed on the right bank to mark the point at which the
measurements were taken.
The location of large organic debris, gravel
bars, and pools were also recorded and referenced to the stakes.
39
These features were plotted on a USGS topographic map (Sitka (C-4)
Quadraxigle, scale = 1:31,680).
During August and September, 1980, cross-sectional profiles were
taken at an average of one every 45.7 m (150 ft) over the same extent
of the stream covered by the theodolite survey, and additional stakes
were used to mark their locations. Individual cross-sections were selected to be representative of each reach of charmel. Also in August
1980, the depth of the thalweg relative to the water surface, stream
width, an4 the distance of the thalweg from the right bank were nieasured at.0.6 m (2 ft) intervals along the 18.6 m (61 ft) reach where
bedload sampling was conducted.
Four cross-sectional profiles were
taken within this reach.
The August, 1981, theodolite survey was restricted to a 259 m
(850 ft) section of the stream, from 1387 to 1646 m upstream (stations 45 + 50 to 54 + 00 of the first survey). Cross-sectional profiles were also taken wherever stakes from the previous study could
be found. The 0.6 ni (2 ft) survey of the pool-riffle study reach was
repeated. Large organic debris, pools, and gravel bars were remapped
for the lower 1646 m of the stream.
Total Suspended Solids and Turbidity
Total suspended:solids (TSS) includes inorganic sediments and
fine to very fine organic matter (0.5 - 1.0 nmi) that is tranported in
suspension in the water column.
TSS saniples were collected during
two 28-hr periods that coincided with portions of three storis during
Fall, 1980. An automatic pumping sampler (Instrentation Specialties
40
Co., Model 1392) was manually activated prior to each expected high
flow event to obtain two subsamples at 30-minute intervals that were
composited in each of the 28 sample bottles held by the machine.
The
ISCO sampler intake was located near the end of a chute, located ap
proximately 1341 m (4500 ft) from the mouth of Trap Bay Creek (see
Figure 18 for precise location).
Samples were filtered through a
4.5 * io8 m glass-fiber filter, oven-dried at 100°C, and analyzed
gravimetrically.
n attpt was made to determine turbidity of the TSS samples with
a Hach Nephalometer.
Unfortunately, a two-hour warm-up period is nec-
essary to stabilize readings on the iristrient, and this was not possible because the electrical source was a gas-powered generator.
The
initial attempts at turbidity analysis without a sufficient warn-up
period yielded inconsistent ad highly variable results.
Bedload Transport Measurements
Bedload samples were collected from the bridges at the pool-riffle
study reach with a hand-held Helley-Sm.ith pressure differential sampler-
during ten storm events during the fall and early winter of 1980.
sampler had a standard 7.6
2
6000
surface area (0.2
ard 1950 cm2 bag.
The
square aperture but was fitted with a
mesh) collection bag instead of the stand-
This larger bag has been shown to reduce clogging
and, thus, improve sampler efficiency (Johnson etal., 1977; Beschta,
1981).
Bedload sampling methods must account for the lateral variations
in transport (Emmett, 1979).
A composite sple was thus obtained by
41
taking subsamples at equally spaced positions along each bridge at the
two riffles.
The upper bridge was marked at ten 0.61
(2 ft) inter-
vals; the lower bridge was marked at eight 0.46 m (1.5 ft) intervals.
Depending on the ambient bedload discharge, the subsamnpling period
ranged from 15 seconds to 1 minute.
A sample was collected from a
given bridge at intervals of from seven to 20 minutes during peak
transport periods, depending upon hOw rapidly the sampler could be
emptied and the water decanted from the samples.
I was able to do this
more quickly as the season progressed, and if sampling was conducted
during the day.
Samples were collected at hourly to bihourly intervals
during the rising and/or falling limb of the hydrograph when bedload
discharge was relatively low.
The lower riffle cross-section was sam-
pled iediately after the upper riffle cross-section had been sampled.
Bedload Sample Analysis
Figure 2 illustrates the bedload sample analysis procedure.
Oven-
dry sample weights were obtained with a Nettler P1210 top-pan balance
(1200 g capacity), accurate to ± .01 g).
Samples were burned at 320°C
f or 24 hours to eliminate organic matter and then reweighed.
Most of
the samples were dried and burned at the Forest Sciences Lab in Juneau;
however, about half of them were burned and weighed at the Forestry
Research Lab, Oregon State Univeristy, Corvallis.
and type of scale were used at both locations.
The same procedure
The seiving and sub-
sequent weighing were done at the Forestry Research Lab.
42
Figure 2.
F10 chart illustrating bedload sample analysis
procedure.
Calculation of Bedload Discharge
Total bedload discharge in kilograms per hour (kg.hr1) were obtamed by dividing the net oven-dry weight of each composite sample by
the number of minutes the sampler was in contact with the bottom, multiplying by 60 min.hr
-1
, and dividing by the fraction of the stream
bottom covered by the sampler orifice (i.e., 7.6
i/bottom width).
Because of the temporal and spatial variability of bedload transport,
the same rate of transport cannot be occurring at all points along
the channel bottom at any given time.
Therefore, the average of sys-
tematic traverses across the channel may be a better indicator of
transport rates than individual measurements (Bagriold, 1977).
However,
this average does not provide an indication of the cross-sectional
spatial variability in transport.
The subsampling procedure does in-
corporate points representative of the varying transport rates across
the channel and also provides an index to the average transport rate
at a given discharge.
Particle size fractions were obtained by seiving (Figure 2) and
were plotted as cumulative distributions of grain size diameter in =
versus percent of sample by weight finer than a given diameter.
Seive
sizes corresponded to the USDA soil-texture classification (Hillel,
1980) and this classification is used throughout this paper.
Inter-
polation from the graphs provided estimates of the distribution characteristics such as D50 and D90.
D50, the median particle size di-
ameter, has been considered to be the siniplest parameter to use in
characterizing the effective grain size dieter (Bagnold, 1977).
D90, the diameter equalled or exceeded by 10% of the particles in the
44
sample, is an index of the largest particle sizes in transport.
E=ett (1981) considers the Helley-Smith sampler to be 100% ef
ficient for particles larger than 0.25
;
therefore, the portion of
each sample larger than this (medit.mi-to coarsesand and gravel) can
be considered to represent total inorganic bedload in transport.
The
portion of material larger than 2.00 .mfl (gravel) excludes any material
which might actually be suspended sediment (Edwards, 1979).
Sediment-discharge rating curves were developed using a power
function of the form:
BLD
where,
aQb
(4)
BLD is bedload discharge in kg.hr1
Q is stream water discharge in m3s1km_2
and a and b are regression coefficients.
This function was also used to develop relationships for total suspended
solids.
Beschta (personal counication) has found that this equation
appears to relate bedload transport to discharge better than a linear
regression equation.
Scatter diagrams also indicated that there was a
curvilinear relationship between discharge and sediment transport.
This
equation is similar in form to that given by Graf (1971) which re-
lates sediment discharge to the actual and "critical" water discharge,
except that no value of "critical ' discharge is assumed.
'Critical'
discharge is that necessary to initiate sediment transport (Graf,
1971).
The same type of equation was used to relate coarse particulate
organic matter transport (CPON), D50, and D90 to water discharge.
Rating curves of this type were developed for all data from all storms
for each of these parameters.
In order to examine the variability of
45
bedload transport relationships between each of the riffles, between
different storm events, and between the rising and falling limbs of the
storm hydrographs, additional comparisons were developed using that
data collected during the time aDd/or from the site of interest.
£11
of the regression equations were tested for significance using an Ftest for goodness-of-fit at alpha-levels of 0.10 and 0.05, and r2
values were computed.
Regression coefficients were tested for signi
ficance using the t-test at alpha-levels of 0.10 and 0.05.
uses a null hypothesis of 'a' or 'b' = 0.
This test
Significance indicates that
the coefficient(s) differ from zero (Neter and Wasserman, 1974).
Organic Matter
Both fine particulate organic matter (FPOM), 0 - 1.0
>1.0
, and CPOM,
, which is in transport near the strebed is collected by the
Helley-Smith sampler.
was made.
No analysis of the size of organic particles
However, all organic matter is referred to as CPOM because
the Helley-Sniith is efficient only for particles in the upper range of
FPOM.
it was not possible to use a furnace èapable of achieving the
standard burning temperture of 550°C for the determination of organic
matter content (American Public Health Association, 1976).
The large
volume of the samples niade it necessary to use a larger oven, and con-
sequently a temperature of only 320°C, for a period of 24 hours.
Sovie
of the organic niatter may not be completely eliminated at this tempera-
ture (Cins, personal counication), but some of the weight loss of
46
obtained using the 550°C temperature may be due to the loss of bound
water from inorganic matter, particularly clay particles (Adams, 1980),
Some bound water
ay even be lost at the 320°C temperature; however I
noted that it frequently appeared that not all organic matter had been
completely eliminated.
Both Adams (1980) and Beschta (personal com-
munication), however, consider the 320°C temperatrue to be adequate for
providIng an index to the relative CPOM content of sediment samples.
The data presented here,therefore,do not represent absolute amounts of
CPOM in transport, but can be considered to represent relative fractions
Work presently being conducted by R.C. Sidle at the Forestry Sciences
Lab in Juneau may provide more information on the accuracy of the analytical procedure used for determining CPOM in this study.
Bed Composition
Bed surface particle size-distributions were estimated from random
samples of surface particles using a procedure similar to that describ
ed by Dunne and Leopold (1978)
Two 1 m2 areas were dslineated at each
of the bridges and in the depositional area between the bridges.
Sur- -
f ace particles were selected by taking a step within the area, reach-
ing down and picking up a particle without looking, and measuring the
particle and replacing it.
each area.
The process was repeated 25 times within
47
V.
RESULTS AND DISCUSSION
Precipitation and Streamf low
Most precipitation during the 1980 study period occurred as long
duration, light to moderate rainfall which did not result in an appreciable rise in streamf low.
Ten storms occurred during this study peri-
od, at which time sediment sampling was conducted; data on precipitation
was available for nine of these storms (Figures 3-10).
Data limitations
make it impossible to characterize storm events as to their relative
intensity; there are no precipitation or streamf low records for this
area other than those collected during this study.
During the early
part of the storm season, it was difficult to predict when rainfall
would result in significant changes in streamflow and bedload transport.
This is reflected in the fact that bedload sampling was con-
ducted .during the 23 and 24 September storms, both of which were rela-
tively low magnitude events (c.f. Figures 3-10).
than 2.0
Peak flows of greater
m3s1 were required to produce appreciable bedload movement
and it was difficult to determine whether rainfall would result in
peak flows of this magnitude.
Peak flows generally had a lag time of about fotir hours follow-
ing the onset of precipitation, depending on the rainfall intensity
and antecedant conditions.
Lag time varied from less than one hour
to more than five hours (Figures 3-10).
Storm flows had a duration
of as much as ten hours and usually lasted more than six hours.
This
made it difficult for me to continue to collect samples over the entire
event.
Therefore, I made an attempt to intensively sample the peak of
48
jtotsi. prsotp. a I5.5 a
.2
-.3
-1.
I
L
.7
-8
2
I-
;
.
3.5k
3.
1.0
0.5
0
1200
1I..00
1600
1800 2000
00 21.00
OO
Tia., hra
Figure 3.
Precipitation intensity and duration and the
resultant hydrograph f or the 23 Septber
1980 storm at Trap Bay Creek, Chichagof Island,
Alaska
49
o
1
-
ti, hr,
Figure 4.. Precipitation intensity and duration and the
resultant hydrograph for the 24 Septiber 1980
storm at Trap Bay Creek, Chichagof Island,
Alaska
50
total
6.0
54.9
5.5
5.0
4.5
2.5
2.0
1.5
.1.0
0.5
1200
Figure 5.
-
1400
1600 iWO 2000 2200 2400
tim., xs
OO OO 0600
Precipitation intensity and duration and the
resultant hydrograph for the 28 September 1980
storm at Trap Bay Creek, Chichagof Island,
Alaska
1200 1400 1603 1800 2000 2203 2400
Precipitation intensity and duration and the resultant hydrograph for the Sept. 30 - 1 Oct. 1980 storm at Trap Bay Creek,
Chfchagof Island, Alaska
1000
tIne, hra
2200 23)0 2403 0203. 0!03 0600 0800
Figure 6.
0
0.5
11.0
Ii
2.0
2.5
total red . = 32.8 na
52
2
4
3
-I
3
3
C
9
7
17.02
idLY?
9.62
62
.i12
65
' .113
2.5
2.0
1.5
1.0
0200
Figure 7.
01.00
0600
0800
1000
1200
1430
tt, T3
1600
1800
2000 2200 2400 0200 0.00
Precipitation intensity and duration aiid the resultant
hydrograph for the 1 October 1980 storm at Trap Bay
Creek, Chichagof Island, Alaska
53
1
total. prcip. s 31.8
2
3
1.
5
6
9
,,
2.5
2.0
0
1.5
1 .0
0.5
0600
Figure 8.
0800
1000
1200
ti, 1
11.00
1600
1800
2000 2200
21.00
00
Precipitation intensity and duration and the resultant hydrograph for the 2 October 1980 storm
at Trap Bay Creek, Chichagof Island, Alaska
54
veage or 2.r thtenit7,
-, V
.Q
..
N
I
N
.2
8
v'
C
'
U.'
0
_g
Figure 9.
V
W%
0
'iqs
v'
0
u
-
Precipitation intensity and duration and the resultant hydrograph for the 5 October 1980 storm
at Trap Bay Creek, Chichagof Island, Alaska
Figure 10.
0
0.5
1.0
'.5
J2.0
r2.5
3.0
3.5
4.0
4.5
1400
1PJa 21k10 22).) 2400
17.0 mrnF
020i)
)4O)
)6i)J
tIm hrj
(1&)i)
IO)t)
totiil jrectj..
I
1200 1OO 1600 1lO) 2000 220J
30.2
6
5
3
0
B
(a
I.
4
PrecipItation Intensity and duration and the resultant hydrograpli
for the 7 October 1980 storm at Trap Bay Creek, Chlchagof Island,
Alaska
1600
total roc1p.
56
the hydrograph, and to take samples at less frequent intervals on
either the rising or falling limb.
Figures 3-10 show the time of duration, average intensity and
2-hour maxim.m intensity of precipitation
and the resultant hydro-
graphs for the nine storms during which bedload was sampled.
Data
were incomplete for the event of 24 September, and there were no data
available on rainfall for the 16 and 18 October events.
The figures
show that the peak of the hydrogaph generally occurs at two to four
hours after the period of most intense rainfall.
Also, because most
rainfall is of light to moderate intensity, the hydrograph often begins to recede while it is still raining
The great point-to-'point variation in precipitation that charac-
terizes this region makes it impossible to compare or relate rainfall
received in one area to that received at a nearby location.
Inghram
(1979) attempted to fill in missing precipitation data for the Kadashan
drainage area, which is west of Trap Bay, using regression relationships developed fram the relatively complete records of the Kadashari
base station and Tenakee Springs.
He found that the method was un-
satisfactory because of the resulting low r2 values, the variability in
the number of days per month for which there was precipitation data,
and the general unreliability of the data.
Inghram also found that
precipitation and discharge comparisons were unsuccessful.
He felt
that there are too many factors involved on a watershed to draw a
useful correlation between precipitation and streamflow.
The Water Resources Atlas for Alaska (1979) lists equations that
can be used to estimate streflow based on selected watershed
57
characteristics.
These equations were used to estimate expected
strea.f low for various return periods and average monthly flows for
August through November (Appendix II).
All events except the 1 Octo-
ber storm appear to have had return periods of less than two years,
based on these equations, whereas the 1 October event had a return
period of between two and five years (Table 3).
The stream gage operated continuously for the entire month of
September.
Mechanical p±oblems resulted in incomplete records for the
other months.
The estimated mean monthly discharge for September from
the equations of the Water Resource Atlas is 1.26 tn3s1 (44.4 cfs).
Actual mean monthly discharge was 0.57
m3s1
(20.3 cfs).
However,
it is not possible to determine whether September, 1980, was relatively
dry, or whether the estimate is too high.
More data on both streamflow and precipitation are needed before
the relative magnitude of flow events can be determined.
The 1 Octo-
ber event was produced by only moderately heavy rainfall (Figure 12),
yet discharge exceeded that of all other events by an order of niagnitude.
The flooding and channel changes accompanying this storm were
much greater than those of any of the other events.
Streamflow at each of the sampling sites was related to the
streamgage readings by regression of the staff gage readings. (one
was located downstream from each of the bridges) against the readings
of the streamgage.
Assu.ing that there was no change in the volume
of water being discharged between the study reach and adjacent to the
streamgage, the area occupied by water at a given stage was determined
at each of the riffles and related to the stage readings at the
58
TABLE 3.
Measured peak flows and estimated recurrence intervals for
Fall, 1980, storms, Trap Bay Creek, Chichagof Island, Alaska
Discharge
Date
3-1
ms
-2
lcii
3-1
Estimated recurrence
interval1, years
fts
23 September
0.046
21.9
<2
23 September
0.147
70.6
<2
28 September
0.417
199.5
<2
30 Sept. - .1 Oct.
0.170
81.2
<2
1 October
1.254
6004
2-5
2 October
0330
158.2
<2
5 October
0.424
203.1
<2
7 October
0.139
66.4
<2
16 October
0.664
3178
<2
18 October
0.167
79.8
<2
)
59
streamgage.
The resulting rating curve used to relate the recorded
stage to voletric discharge is illustrated in Figure 11.
Bed Coposition
The streambed of Trap Bay Creek consists of particles derived
from igneous bedrock and ablation till.
Past glaciation has acceler-
ated the breakdown of the relatively resistant bedrock, thus increas-'
ing the proportion of bed material in smaller size classes.
There is
a wide range of particle sizes, the larger of which may exceed 10 cm
in diameter.
Larger particles are usually angular but those derived
from glacial deposits are often somewhat rounded.
Much of the sur-
ficial material is made up of small to medit-sized particles which
range frog less than one to several
material is underlain by alluvi
in diameter.
This non-cohesive
and ablation till, both of which may
contain large amounts of colloidal-sized particles and silt.
Results of surface-particle size distribution sampling are presented in Table 4 and Figure 12.
There does not appear to be a sig-
nificant difference between the riffles and the pool except that the
median particle size in the pool and lower riffle was slightly larger
than that in the upper riffle, with that of the lower riffle being the
largest.
Time of spling must be considered in interpreting these
results.
Sampling was conducted in late August; the armor layer could
have been well-developed over the whole reach.
Differences in bed
composition may not have been evident in surface material because of
the churning of the gravels by pink salmon, although this reach is not
60
120
105
90
45.
30
15
0
OJ0
Figure 11.
0.75'
1.00
1.25
Stage, ft.
1 e50
1.75.
2.00
Stagedischarge1rating curve for Trap Bay Creek,
Chicagof Island, Alaska
1Toconvert f9m feet to meters, multiply by 0.304S, to convert
from cfs to m si multiply by 0.028.
61
heavily used by fish.
It is evident, however, that the largest per-
centage of surface particles (about 50%) is composed of coarse sand,
with most of the remainder (about 45%) being gravel.
Channel Morphology
The thalweg survey was only repeated for a 259 m (850 ft) section
of channel in 1981; both the 1980 and 1981 surveys were conducted in
mid-August.
Results are presented in Figures 13 and 14.
channel gradient in 1980 was 0.25%.
Overall
Gradient decreased from 0.24% in
the resurveyed section to 0.17% from 1980 to 1981, while stream width
rained about the same.
The resurvey of the pool-riffle study reach,
where measurements were made at 0.61 m (2 ft) intervals, showed that
no major changes in bankfull width had occurred from 1980 to 1981
(Figure 14).
However, gradient in this section nearly doubled, going
from 0.41% to 0.81%, as a result of aggradation at the upstream end
and degradation at the downstream end.
The greatest changes in thalweg elevation atid location took place
upstream from a chute (Figure 13), except for some aggradation and a
shift of the thalweg towards the right (looking upstream) near the
lower end of the chute (Station 45 + 00).
This chute appears to be
relatively stable, probably because of the material which underlies
it.
Boulder-sized material becomes apparent at about 122 m (400 ft)
from the upstream end of the chute, and increases in size and abundance in the upstream direction.
The tidal influence zone reaches as
far up as the chute, but does not appear to reach the stream gage.
The lack of extreie fluctuations due to tides may contribute to
62
TABLE 4.
Class
Size
Size Distributions of Particles Selected at Random From
the Armor Layer of the Study Reach Trap Bay Creek,
Chicagof Island, Alaska
no0 of
particles
in class
Z of
total
cumulative
D95
% finer
lower riffle
0.25
0.25
0.50
1.00
2.00
1
2
0
3
11
6
6
2
22
12
24
48
30
42
4.00
5
10
90
8
2,43
0.44
4.30
1.66
0.44
4.40
2q00
039
480
upper riffle
0.25
O.25
0050
1.00
2.00
4.00
1
2
3
6
10
17
15
4
20
34
30
0
2
8
28
62
92
8
pool
0.25
0.25
2
1
4
2
0
4
0.50
12
10
24
20
30
17
34
16
1q00
2,00
4q00
8
6
50
84
63
100
qo
10
-
30
0
30
10
5
4
%
a
C
c
O.3
c.
0.1
Graiz size diater. cm
Figure 12.
Size distribution curves developed from random samples of surface particles in the study reach, Trap
Bay Creek, Chichagof Island, Alaska
0
.
43
.
H
5
0
'4
610
16+00 20+00
iq
Figure 13.
i
t
0.17
0.24
0.25
,
0.81
0.41
I
I-
*
distance
depth
1
0,5
1981
1980
1 586
$
52+00
fart f'nr nTnlanatlnn nf sttinn aqtablishmant.
of low tIde mark, 22 July 190 (see text).
Stream width and distance of thalweg from leIt bank, (upper graph) and
thalweg depth2of Trap Bay Creek, Chichagof I1and, Alaska
*
854
in
32400
36+00 40+00
44400 48400
I
I
a
5
iO98 122p, ,1 342
976
144
stream gradient, %
1981
1980
732
I
or distance from low tide mark
24+00 28+00
study reach
resurveyod reao
total distance surveyed
12+00
Station, ft'
-
I
I
I
Control
Creek
1593
1981
- -
a
1597
dldtrino6
organ a
debria
tree
plan view
-
-
gravel
1609
,depos1t
Creek
Control
trout ntrwm mouth, m
1600
1606
1603
a
I
upper bridge
I
19tO
1612
Thalveg profile and planimetric map of the poof-riffle study reach
along Trap Bay Creek, Chicagof Island, Alaska (surveyed in September
1980, and August 1981)
lower bridge
I
It
1590
Figure 14.
10
a5
a
a)
1). 0
4.)
ri
66
stability of the chute.
The map of stream features (Figures 15-18) shows that the chute
is less subject to bank sloughing and inputs of large organic debris
than are upstream or downstream reaches.
a result of its relative stability.
This may be both a cause and
Cross-sectional profiles taken
in this reach (not shown) indicate little net change had taken place
from 1980 to 1981.
Very few salmon spawn in this reach.
This may be
because of the difference in streambed material composition, but it
may also be due to lack of cover in the form of undercut banks and
large organic debris (Eubbardt9 Alaska Dept0 of Fish and Game, personal
counication)
The lower reaches of the channel are subject to a great deal of
morphometric change.
Bank-sloughing, tree-tipping, and shifting of
gravels were widespread within the tidal influence zone (Figures 1517).
The meander just above the mouth of the chaimel has shifted
from west to east several times in the past based on an analysis of
aerial photographs, and is actively cutting its banks0
This section
of channel has also been subj.ected to much human activity: trampling
of the banks, removal of large organic debris, and construction of two
fish weirs.
The first weir was constructed in 1979 and was destroyed
by high flows before it was completed.
ed in July of 1980.
The second weir was construct-
Some of the changes in this section of the chan-
nel are probably due directly to human activity.
The resurvey of the poolriffle study reach showed that the thalweg elevation has increased and the channel has widened downstream
from the upper riffle (Figure 14).
Downstream from Station 52 + 62,
67
t.
ci
Jtc
e.Jnt
A
ett
,
i
;+
t4.
c.e*t
&CtTh;Aq
crn'
50-4+.,
%1ker%!15 S
c;
ètc.
or
Lt
3n
Figure 15.
Planimetric map of the lowest 923 m (2700 ft) of Trap
Bay Creek showing changes from August 1980 to August
1981
68
i
*f
,t *IL.
eM
'..
,e.t óoi.iX%
,
1%
a
. \e4.t.
L_
o.
i3U3 4M3 'Flt
e.
'rA4.
I%)fl.
'
c.nà
tG
W;
'So
04
Th
tsar.
rm
urrm Q1Y
'
Keu
£14.
s
.t.
.'t'
I
t'3.3 .L
Figure 16.
Planimetric map of Trap Bay Creek frorn 823 m (2700 ft)
from stream mouth to 1219 m (4000 ft) from stream
mouth showing changes from August 1980 to August 1981
69
I4
Ut 3#rn.
-tf.
br
Q%43 t
Figure 17.
Planimetric map of Trap Bay Creek from 1539 m (5050 ft)
from stream mouth to 1844 m (6050 ft) from stream mouth
showing changes from August 1980 to August 1981
70
4O4G0
AX
4
4ôed wi
Q?
&' 'rf. b
rl
0.
-
ifl31
t3
tdøJ
er.
;Aue.
3i.3.s
ce
ewv trr
=
t. 50 4t.
trnAr
GrtJ hor
mi1 Grf
0
Aa&, %qS
od&I'%
14 C.Ø
Figure 18.
Planimetric map of Trap Bay Creek from 1219 m from stream
mouth to 1539 m (5050 ft) from stream mouth showing changes
from August 1980 to August 1981
71
the true left (tttrue left" is relative to an observer looking down-
stream) bank is being undercut and the thaiweg has shifted towards
this bank.
riffle.
Little change in thalweg location occurred in the upper
Deposition has occurred in the lower section of the pool be-
tween the riffles.
Scouring occurred in another pool located 30.5 m
(100 ft) downstream from the pool-riffle study reach and the thalweg
elevation decreased by nearly 0.3 m (1 ft).
tIpstrei from Station 52 + 68 for about 61 m (20 ft), the thalweg has tended to shift towards the middle of the stream (Figure 14).
Gravels have begun to acci.ulate at what was the major outlet of Control Creek and debris has acci.ulated at the true left ends of the
two fallen trees near the upper bridge (Figure 18).
The debris tends
to channel flow into the middle of the stream while the buildup of
gravel at the foot of the upper bridge tends to channel it back towards the true left bank.
The cross-sectional profiles taken in the primary study reach
(Figure 14) show that, although there has been an increase in the thalweg elevation, there has actually been net scour (see also Table 8).
Although more material was transported out of this reach than was
transported into it, there was a tendency for the maximum depth of
the channel to increase because the channel bottom became more even.
The survey of 1980 included 1951 m (6400 ft) of channel.
The
gradient of the channel from Station 53 + 00 to Station 64 + 00 was
approxImately 0.80%.
There was much more large organic debris in the
upper portions of the channel (Figure 18) and it was difficult to find
any pool-riffle sequences which were not a result of the pressure of
72
large organics.
Much of the organic material in the stream above the
study reach was tree-sized and relatively stable.
Smaller debris can
be trapped against fallen trees and debris build-ups resulted in localized scouring and deposition in many locations.
Large organics are a dominant influence on channel iorpho logy.
Tree-sized material, appears to generally remain where it falls for
relatively long periods of time, except in the extreme lower reaches
of the channel.
Localized scour and deposition take place where there
are build-ups of debris, disrupting any natural pool-riffle sequences
in the upper reaches.
Trees in the channel do not appear to block
fish passage but, instead, serve as cover for spawning adults.
Both
large organics and undercut baxks are important as cover for juvenile
salmon (Swanson, 1980).
Width-depth rations were computed for all of the cross-sections
which were resurvyed in 1981 (Table 5).
In general, the ratios
increased from 1980 to 1981, indicating that the channel is widening
and/or becoming more shallow.
The average stream width increased from
13.0 to 13.9 m (42,6 to 45.6 ft) and the average depth decreased from
0.91 to 0.82 m (2.99 to 2.69 ft) from 1980 to 1981 over the 914 m
(3000 ft) section in which cross-sections were measured.
widening and aggradation appear to be occurring.
Thus, both
The width and depth
of cross-sections within the chute remained relatively constant from
1980 to 1981, however, again indicating that it is relatively stable.
The morphological changes that took place in Trap Bay Creek from
1980 to 1981 indicate that it is an active channel undergoing aggradation and widening.
Lithology is one factor determining the morphology
.:
width,
pi
Ii
0
Stothrn 5) + 15 ( 1620 * frog low til. isork)
... .
width,
-4...
1,
r lo Ud. piork)
10
19
17g5
10
174
debrii
organS..
till
1001W
- ---21 Awl. 1961
14 Aug. 19I0
frous low ti. .iork)
10
frc low tida .sirk)
width,
- sipj.r rIft 1.
I
t.t1osa 52 4 79 ( 1609
2.5
pi
S
.14th, ii
station 52 4 40 (1597
2.5
Net channel changes from 14 August 1980 to 21 August 1981 within
the pool-riffle study reach of Trap Bay Creek, Chichiagof Island,
Alaska
25
3t4Uou 52 * So ( i600
25
(rca lois tid. i.siik)
5tiiLio 52 * 18 (1590.4
- lotior rifil.
13
Figure 19.
I)
9
width,
2,5
74
TABLE 5.
Station
Width-depth Ratios for Selected Stations1 in 1980
and 1981) Trap Bay Creek3 Chichagof Island, Alaska
Width,
1980
Depths
W/D Ratio
Width,
1981
Depth,
W/D Ratio
m
54
53
52
52
52
52
52
51
50
49
48
46
44
43
42
37
35
34
2J
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
00
15
79
50
40
18
00
00
00
40
00
05
90
50
00
50
00
00
50
135
172
12.8
11.1
10.2
0.79
0.97
1.13
0.84
0.76
99
053
10.4
0.58
137
084
9.8
12.5
13.7
0.87
0.93
140
141
165
0.84
12.5
14.5
10.5
11.8
18,9
0.70
0.99
0.99
0.70
0.23
13.0
041
091
069
076
037
24OOi3T4
average
,7Y
17.0
136
176
17.7
13.4
11.7
10.9
10.7
11.4
13.3
134
186
18.0
16,4
112
13.4
20.0
1L8
13.7
11.3
12.5
0.68
0.79
0.91
0.82
0.52
0.61
0.46
0.55
0.76
0.85
140
139
084
14.,1
076
45.0
17.9
14.6
10.6
16e5
037
052
168
121
241
1.8
1.68
0.88
0.76
0.15
0.38
13.9
0 ..82
167
184
82.7
32.6
(p77
125
15.3
11.6
I
19.8
22.3
14.7
140
21.1
17.5
258
25.0
14.8
14.6
28.8
.049
166
184
450
24.2
9.1
13.1
15.9
158.0
41.6
5.
1Stations decrease in value in the downstream direction; see
text (pp. 43-44) for explanation of station establishment
ere
(dcf
do
).
i4
/
1
II.
75
of Trap Bay Creek as is indicated by the presence of the relatively
stable chute.
The tides are important in the lower reaches of the
channel, especially where there are few trees lining the streambanks.
Channel gradient appears to be a function of both lithology and the
tidal influence zone.
Large organic material, especially fallen trees, interact with
streamflow and sediment transport in determining chatmel morphology
above the tidal influence zone.
Smaller organics (branches, twigs,
bark, and leaves) are also important in that they can lodge against
fallen trees and gravel deposits, causing local deflection of flow.
Organic material is also important in providing cover for salmonids.
Total Suspended Solids
Problems involved in processing the total suspended solids (TSS)
samples precluded a fully accurate analysis of suspended sediment
transport in Trap Bay Creek.
Electricity to power the filtering ap-
paratus was not always available, so the entire set of samples usually
had to be transported to the FSL in Jmeau.
Here, additional probls-
arose because the filter discs being used were not consistent in
weight and preweighing in the field was impossible.
Trap Bay Creek
has a relatively low suspended solids concentration even at peak
flows.
Filter disc weights varied by as much as a gram, which is three
orders of magnitude greater than the total mass of most of the samples.
Preweighing of filter discs in the lab before transport to the
field gave reasonable accuracy, but was tiie-constziing and dependant
on a lack of
.ix-ups when the filter discs were used.
Data for two
76
sets of samples covering portions of three storm events were considered
reliable enough for interpretive use.
The data presented in Table 6 show that TSS concentrations ranged
from 0 to 86
mgl1
during flows ranging from O.O5 to 1.26 cubic me
ters per second per square kilometer
(m3s112).
The naturally low
suspended sediment regime is likely a result of the relatively high
erosional resistivity of the rocks in the watershed, low intensity rainfall, rapid soil infiltration rates and subsequent lack of surface run
off.
Suspended sediment loads of nonglacial streams in southeast Alas-
ka are generally extremely low (Schmiege etal., 1974).
component of many high re1ief
The bedload
glacial-form watershed streams is large
and this limits the lasting influence of suspended sediment in the
streambed.
Trap Bay Creek represents a d±fferent type of stream than
those streams with high suspended to bedload ratios which have been
studied in the southern U.S. It appears that most sediment transport
occurs as bedload here, and that suspended sediment is relatively unimportant in this undisturbed old growth system.
Sediment rating curves were developed for TSS using the equation
(4).
r2 values ranged from 0.18 to 0.999.
port at 0.06
m3s1i2
Prediction of .TSS trans-
the average annual flow according to the
Water Resources Atlas for Alaska (1979), ranged from 2.1 to 26.1
mg.l
(Table 6).
A plot of TSS versus discharge (Figure 20) showed
that, although there is a slight hysteresis effect evident for the
24-23 September data, the opposite is true for the 1 October data.
A
hysteresis effect is coonly seen in TSS data collected from streams
in W. Oregon (Milhous and Klingeman, 1973; Beschta, 1981; Edwards,
1979).
Data from one event at Trap Bay Creek do not provide a basis
77
TABLE 6.
Summary of Total Suspended Solids (TSS) Data Collected
From Trap Bay Creek, Chichagof Island, Alaska, During
Two Storms in the Fall of 1980.
24-25 September 1980
Ti.me,
hrs.
30 September - 1 October 1980
Discharge
TSS,
m3sllcxn2
mg 1-1
0.07
0.07
0 08
0010
10
1800
1900
2000
2100
2200
2300
2400
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
016
0.16
0.13
0.11
0.10
0.09
0.08
0.07
0.06
0.07
0.06
0.06
0.06
0.05
0
5
28
32
27
0
26
1
1
32
2
0
0
3
0
0
2
Time,
hrs.
Discharge
m3s 12
mg 11
2300
2400
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1200
1300
1400
1500
1600
1700
0.13
0.13
0.11
0 11
0.11
0.11
0.11
0.12
0.12
0.16
0 25
0.39
48
42
44
33
25
28
21
0.71
1.09
1.26
1.06
0.71
0.49
61
63
TSS,
34
40
18
51
49
86
72
52
38
TSS rating curves1
Predicted TSS transport
for the average annual
discharge. Q = 0.062
all data
TSS = 96.5 Q087 r2
24-25 September : TSS = 2067.2 Q236r2
rising limb
TSS - 1412.7 Q2.05r2
falling limb
TSS
3366.4 Q2.05r2
30 Sept.-1 Oct.
rising limb
falling limb
TSS = 65.0 Q
TSS = 64.4
TSS=68.6Q
033
r
r
'
0.41
8.25
0.34
0.67
0.18
2,76
4.43
2.08
0.59
0.46
25.54
26.10
6.79
r=1.00
See text for explanation of rating curve development.
Average annual discharge computed using an equation from the
Water Resources Atlas for Alaska (1979).
78
90
/
'I
///1
,,/1II
80
70
-
60
-,
A:-,
A
P3o Sept.-',"
f "I
/1 Oct. -I
hO
p
30
p
I
I
20
24-25 Sept.
0.5
Figure 20.
0.5
0.65
Discharge,
Q.I5
0.5
311.2
1.d5
Total suspended solids (TSS) versus stream discharge for
two time periods during the Fall of 1980, Trap Bay Creek,
Chichagof Island, Alaska
79
for concluding that TSS hystersis does or does not occur.
The 1 Octo-
ber event may have resulted in TSS transport unlike that which occurs
during more frequent events.
The fraction of suspended load that is made up of organics
is variable and can compose all or none of a given sample (Figures 21
and 22).
The proportion of organics does not appear to depend on dis-
charge or on the total amount of TSS in transport.
It is interesting
that, during the 24-25 September event, organic material makes up a
relatively small portion of TSS on the rising limb of the hydrograph,
but TSS samples on the falling limb were frequently coiposed entirely
of organic material.
In contrast, organics initially comprised 20 to
50 percent of TSS at the beginning of the 1 October event, but are
essentially absent from TSS samples collected during the peak and f alling limb.
This may be another indication that TSS transport during
this event was not characteristic of the stream during relatively normal flows.
Bedload Discharge
At the lower riffle of the pool-riffle study reach, bedload discharge, including organic material, ranged frog 3.9 to 4200 kg.hr1.
At the upper riffle, bedload discharge ranged from 15 to 4400 kg.hr4.
The lowest measured transport rate occurred during the first storm of
the season, 23 September, at the lower: riffle, and during the 7 October storm at the upper riffle.
The greatest measured transport rate
occurred during the 1 October event at both sites but the time of peak
transport differed.
Peak bedload transport occurred nearly coincident
80
100:.
.1 7
.o10
80
1
S
.125
E
0
.4.
o40
S
0'
20
.075
1
.050
1900
Figure 21.
2400
000
1000
Ti,hrsT
Total suspended
organics (TSO),
time during the
Trap Bay Creek,
1.500
2000
scuds (TSS), total suspended
and stream discharge (Q) over
storm of 24-25 September 1980,
Chichagof Island, Alaska
81
0
2300
Figure 22.
0400-
0900
Tixne, hrs.
1300
1800
Total suspended (TSS), total suspended
organics (TSO), and stream discharge (Q)
over tine during the storm of 30 Sept.1 Oct. 1980, Trap Bay Creek, Chichagof
Island, Alaska
82
with the peak of the hydrograph at the lower riffle, but showed wide
variability at the upper riffle (Figures 23-33)
Peak transport rates occurred nearly coincident with peak discharge during the 30 Sept. - 1 Oct0 and 1 Oct
incomplete for the other eight events.
events0
The data are
Transport rates at the upper
riffle exceeded those at the lower riffle during all events except
that of 1 Oct., the latter part of the 2 Oct. event, and the 7 Oct
event.
Apparently, sediment was transported past the upper riffle
and deposited in the pool during lesser, more frequent events.
An
event of sufficient magnitude was then required to dislodge this ma-
terial and transport it past th
lower riffle
Once this material had
been dislodged, transport past the lower riffle was accomplished by
lesser magnitude events until all available sediment had been. trans-
ported and the depositional area becaie essentially rearmored.
Langbein and Leopold (1960) theorized that a gravel riffle is
a
ezpression of a kinematic wave, and that it requires repeated flows
of sufficient magtiitude to transport material from one zone of concentration to the next.
Bay Creek.
This theory could explain what happened in Trap
Lesser events are capable of transporting material to and
past the upper riffle, which represents a "zone of concentration."
A
series of greater magnitude events was necessary to transport this material past the lower riffle to the next zone of concentration.
The average change in sediment storage in the pool-riffle study
reach was computed from estimates of net change in cross-sectional
area from 1980 to 1981 in the cross-sectional profiles.
shown in Table 7.
Results are
The average volume of sediment transported during
83
V
'V
/
C
0
C
'4'V
w
0900
1000
1100
1200
1300
ILCO
tthe, ?3
5OO
1600
1700
1$CO
1900
2000
Figure 23.
Discharge (Q), bedload (BLD) and organic
bedload (BLDOrg) in transport, and average
particle diamter of bedload sample CD50) over
time for the 30 September 1980 storm at
Trap Bay Creek, Chichagof Island, Alaska
NOTE:
o indicates samples from the upper riffle
x indicates samples from the lower riffle
84
C
-4
C.
I.
a
C
=
C
C
a
0
1500
1600
1700
1800
1900
t, s
2000
2100
C0 2300 2h0 C00 0200
Figure 24.
Discharge (Q), bedload (BLD) and organic
bedload (BLDOr) in transport, and average
particle diameter of bedload samples (D50)
over time for the 24 September 1980 storm
at Trap Bay Creek, Chichagof Island, Alaska
NOTE:
o indicates samples from the upper riffle
x indicates samples from the lower riffle
85
1OO 1600 1703
Figure 25.
NOTE:
13C0
t, 2000
kr 2100 2200
1900
Z.1300
2LA
Discharge (Q), bedload (BLD) arid organic
bedload (BLDorg) in transport, and average
particle diameter of bedload sample CD50)
over time for the 28 Sept. 1980 storm at
Trap Bay Creek, Chichagof Island, Alaska
o indicates samples from the upper riffle
x indicates samples from the lower riffle
86
0
as
0
I..
C.
1803
19Cc 2000
21 GQ
OO 2QQ ZhQ O1CO
t9 b5
Q O3O
O5O
Figure 26.
Discharge (Q), bedload (BLD) and organic
bedload (BLD0rg) in transport, and average
particle diameter of bedload sample (D50)
over time for the 30 Sept. - 1 Oct. 1980
storm at Trap Bay Creek, Chichagof Island,
Alaska
NOTE:
o indicates samples from the upper riffle
x indicates samples from the lower riffle
J
ec.i
a
USLI
t__L__L__I1 i.._..L_.i_
uosi o( vo,(
000
O.O061
091 VL1 091 (ci
I
I
oc.
Figure 27.
I
a
&
I
I
IM.c
ee
I
't19
000
I
I
C'5L1
I
a
I
0.9
I I
oct ccii
I'a CC1 en ou 001 06
I
c.3c1
0L
I
e.oa
-
I
c
I
-
c
o'
oc ox 01
00
0
a
101
___1 .iLI- 0i ftla. a
.(
a..
a
0
-
R
0
S
S
Discharge (Q)I bedload (BLD) and inorganic bedload (BLDOrR) in transport, and average
particle diameter of bedload sample '(D50) over time for tI'e 1 October 1980 storm at
Trap Bay Creeks Chichagof Island, Alaska
88
a
C.af
qa
C
C
C
170
IEGO
1900 2GC 2100
-,
22
I
2300
21.0
tine, br
Figure 28. Discharge (Q), bedload (BLD) and organic
bedload (BLDorg) in transport, and average particle diameter of bedload sample
(D50) over time for the 2 October 1980
stcr at Trap 3ay Creek, Chichagof Island,
Alaska
NOTE:
o indicates samples frog the upper riffle
x indicates samples from the lower riffle
I
1.
I
6
C
igure 29.
-,
2
I
I)
1
'I
II ii
0.5
J_iJ
0
250
I I
1.0
15
I
3.5
4.0
4.5
5.0
5.5
6.0
1750 2000 225() 2500 2750 3000
1
I
I
I
1 I 1
11
3.0
1500
I
2.5
1250
2.0
1000
I7501%I
I
,I
oo:
I
Discharge (Q), bedload (BLD) and organic bedload (BLDorg) in transport, and average particle diameter of bedload sample (D50) over time
for 5October storm at Trap Bay Creek, Chichagof Island, Alaska
I
bID1 kg4ir1
o indicates samples from the upper riffle
.x indicates samples from the lower riffle
BLDorg kg.hr1
0
7i
0
25
NOTE:
90
1000
1100
1200 1OO 1LO 15Cc 1éO
7OV
s
Figure 30.
Discharge (Q), bedload (BLD) and
organic bedload (BLD0 ). in transport,
and average particle ä.ameter of bedload sample (D50) over time for the
7 October storm at Trap Bay Creek,
Chichagof Island, Alaska
NOTE:
o indicates samples from the upper riffle
x indicates samples frog the lower riffle
91
0900
1000
1100
1200 10C lh.00
ti.,
:s
1500
16C0
Figure 31.
Discharge (Q), bedload (BLD) and organic bed-.
load (BLDorg) in transport, and average particle diameter of bedload sample CD50) over time
for the 16 October storm at Trap Bay Creek,
Chichagof Island, Alaska
NOTE:
o indicates samples from the upper riffle
x indicates samples from the lower riffle
92
C
10
0
0
I,
4,
1900 2000 2100 2200 2300 2h0O
Figure 32.
Discharge (Q), bedload (BLD)
and organic bedload (BLD0rg)
in transport, and average particle diameter of bedload sample CD50) over time for the
18 October storm at Trap Bay
Creek, Chichagof Island, Alaska
NOTE:
o indicates samples from the upper
riffle
x indicates samples from the lower
riffle
93
each storm was computed at each sampling site by multiplying the average transport rate by the number of hours that significant bedload
transport occurred.
A su=ation of these voltmies for all storms gave
an estimate of the volume .of sediment transported past each of the riffles (Table 8).
The estimate based on cross-sectional areas was 10.8
tonnes (11.9 tons) of sediment transported out of the study reach from
1980 to 1981.
Based on the change in each cross-section, 20.4 tones
(22.4 tons) of this material was scoured and 9.6 tonnes (10.6 tones)
was deposited.
These figures are comparable to the estimates of 11.4
totnes (12.5 tons) of sediment transport past the upper riffle and
20.7 tonnes (22.8 tons) of sediment transport past the lower riffle.
More sediment was transported past the lower riffle, which agrees
with the net negative chaige in cross-sectional area here, and the
positive change in area at the upper riffle (Table 7).
While these
figures are only rough estimates, they indicate that Helley-Smith samples
ay provide a reasonable estimate of bedload sediment transport
if the changes in channel cross-sectional area are considered to represent the change in bedload sediment storage in the study reach.
Plots of bedload discharge > 0.25
no hysteresis was obvious.
(not shown) indicated that
These plots also showed that, at the lov-
er riffle, the bedload transport at a given discharge during the 2 October and 5 October events was greater than it was during any other
events.
These wo moderate events appear to have transported most of
the sediment that was made available by the 1 October event past the
lover riffle.
Bedload discharge during the 16 October event at a
given discharge was less than that of the 1 October event even though
94
TABLE 7.
Change in Sediment Storage Computed from Cross-Sectional
Area Changes from September, 1980, to August, 1981, for
Trap Bay Creek, Chichagof Island, Alaska
Area,
m (ft)
Station
1
Length,
m(ft)
18(19.3)
scour
0.9(10.0)
scour
1.3(13.5)
0.01(0.05)
scour
0.7(7.0)
1.4(15.3)
scour
1,2(130)
2.7(9.0)
-2.36(83.25)
7.6(270.0.)
6.1(20.0)
-7.61(269.0)
0 . 02 (1. 0)
- 3.2(112.0)
52 + 40
fill
m3(ft3)
+ 2.6(90.0)
52 + 18
fill
Net Chaxige2
m3(ft3)
- 4.9(173.3)
52 + 00
fill
Vol.mie,
+3.74(132.0)
4.9(16.0)
+6.9(244.0)
7.2(2535)
52 + 50
-276(97.5)
5.9(19.5)
fill
07(8.0)
+ 44(156.0)
scour
0.6(6.75)
-6.2(219.4)
fill
0.9(9.25)
5? + 79
9.9(32.5)
+2.25(79.6)
4-8.5(299.0)
Total
-674(238.2)
* 1.6 tonnes/m3
10,8 tones
Total scour = 20.4 tomies
1
Total fill = 9.6 tonnes
See text for explanation of station estab1ishnent (pp. 43-44)
2+ indicates net fill; - indicates net scour.
3Based on Hillel (1980).
3
95
TA3LE 8.
Approximate Amount of Sediment Transported by Fall, 1980,
Storms, Trap Bay Creek, Chichagof Island, Alaska2
Storm date
Ave. trazisport
rate, kg.hrl
Time,
hrs
Sediment transported
tonnes
tons
upper riffle
23 Sept.
33.0
4.0
0.13
0.15
24 Sept.
188.8
3.5
0.66
0.73
28 Sept.
198.6
2.5
0.50
0.55
71.1
4.25
0.30
0.33
1 Oct.
962,3
5075
5.53
6.10
2 Oct.
792.4
2.0
1.59
1.75
5 Oct.
397.4
4.5
1.79
1.97
7 Oct.
35.6
4.0
0.14
0.16
16 Oct.
322.4
0.73
Total 11.37
0.80
12.53
30 Sept.- 1 Oct.
2.25
23 Sept.
5.2
lower riffle
4.0
0.02
0.02
24 Sept.
34.7
3.5
0.12
0.13
28 Sept.
47.3
2.5
0.12
0.13
50.3
4.25
0.22
0.24
1 Oct.
2402.8
5.75
13.82
15.23
2 Oct.
430.8
2.0
0.86
0.95
5 Oct.
839.2
4.5
3.78
4.16
7 Oct.
319.1
4.0
1.28
1.41
16 Oct.
229.1
2.25
0.52
Total 20.72
0.57
22.84
30 Sept.- 1 Oct.
1Time that significant bedload transport occurred.
determined by bedload sampling with a Helley-Smith pressuredifferential sampler.
96
the peak flow of the 16 October storm was greater than that of the 2
and 5 October events.
At the upper riffle, transport rates early in the storm season
were comparable to transport rates at a given discharge following the
1 October storm.
Transport rates increased with storm magnitude.
However, the 1 October event did not produce au increase in transport
comparable to the increase at the lower riffle except for the slug of
material at the end of the storm (Figure 28).
Transport rates during
the 7 and 16 October storms were comparable to those of the 1 October storm at discharges of one-fourth and one-third, respectively,
that of the 1 October eveut0
This could have been a new wave of ma-
terial moving into the pool to replace the material that was transport-
ed out of it by the 1 October storm.
Bedload, CPOM, D50, and D90 were regressed against stream discharge using the aforemeutioned power equation (4).
Equations using all
data from all storms combined are suinmarizedin Table 9.
The data
were separated according to the site from which the samples were collected and regressions were developed for each of the riffles (Table
10).
In Figure 33, the graphs for all significant relationships are
presented; Figure 34 depicts the significant relationships for each
riffle.
Equations are depicted over the range of discharge observed.
Separation of the data by sampling site indicated that there was
a better correlation between discharge and bedload transport at the
lower riffle than at the upper riffle.
Correlation coefficients for
relationships for the two bedload size fractions and CPOM were higher
when only data from the lower riffle were used than when equations were
97
TABLE 9.
Relationships Between Bedload (BLD) Tratisport (BLD > 0.25
and BLD 2.0 >
in kg.hr), Coarse Particulate Organic Mat-
ter (CPOM in kg.hrl), Two Particle Dieters (D50 and D90
in mm) and Streaflow for the 1980 Fall Storm Season at Trap
Bay Creek, Chichagof Island, Alaska
= 132 (all storms)
Equation
(BLD > 0.25 mm) - 1534.12 Q1613
2
ti ,Z
F'
a
b
0.69
**
**
**
r
mm) -
877.47 Q1719
66
*
*
*
(CPOM)
-
89.15 Q1857
0.71
**
**
**
(D50)
=
(D90)
=
(BLD > 2.0
2.55Q
0.158
13.71 Q015°
0.04
*
0.7
*
indicates sigificace at the 90Z level of probability;
** indicates significance at the 95% level of probability.
2significace indicates that the regression coefficient differs
froi zero.
98
TABLE 10.
Relationships Between Bedload (BLD) Transport (BLD > 0.25 mm
and BLD > 20 mm in kg.hr1), Coarse Particulate Organic
Matter (CPOM in kghr1), Two Particle Diameters (D50 and
) and Streamfiow (Q) for Each Sanipling Site on
D90 in
Trap Bay Creek, Chichagof Island, Alaska, for the Fall,
1980 Storm Season
upper riffle (n
66)
t ,
Equation
r
2
F
1
a
b
(BLD > 0.25
) - 1021 Q125
0.58
**
*
(BLD > 2.0
) -
555 Q130
0.56
*
*
*
*
(CPOM)
- 70.3
Q177
0.66
(D50)
- 2.12
Q°"5
0
126 Q°'°6
(D90)
*
0.01
lower riffle (ii
66)
(BLD > 0.25 =) - 2267 Q198
0.82
**
**
*
(BLD > 20
) - 1379 Q115
0078
**
*
*
112 Q195
0.77
**
*
*
(CPOM)
-
(D50)
- 3.06
Q°32
0,13
(D90)
- 16.5
Q°3°
0.19
*
indicates significance at the 90% level of probability;
** indicates significance at the.95Z level of probability.
2significance indicates that the regression coefficient differs
from zero.
I.,
0.4 0.6 1.0
0.66
I-
2
3
6:
43-
1.
0.04 0.06
H
H
3- 0
4.p
6:
0.2 0.4 0.6
Q. m3s-1km2
0.1
1.0
(cpoI4) = 89.2 Q16
r2 0.91
Bedload (BLD), coarse particulate organic matter (CPOM), and particle dlor D90) vs. streamfiow (Q) relationships for all data (n = 132)
collected from Trap Bay Creek, Chlchagof Island, Alaska, during the Fall,
1980, storm season
anieter (D5
0.2
r2
(BLD > 2.0 mm)
wrY 5
(rnD >0.25 im) - 153at.i Q1161
0.69
0.02 0.04 0.06 0.1
-
Figure 33.
3-.
4-
6:
2..
6:
1-
2-
0
-.3
H
6-.
.4
a 0.56
02
I
(api)
70.3 Qh7?
= 0.66
Q.4 0.6 1.0
)
0.04 0.06
0.2 0.4 0.6
Q m3s1hr2
0.1
Lover fliff 1.
1.0
Bedload (BLD), coarse particulate organic matter (CPOH), vs. streamfiow (Q)
relationships for data collected from each sampling site (n = 66) on Trap
Bay Creek, Chfchagof, Alaska, during the Fall, 1980, storm season
Q, m3ekia2
0.02 0.04 0.06 0.1
(BLD> ?Q
= 555.2 QL.,O
(BLO ) 0.25
1021.3 Ql.25
r2 0.58
Figure 34.
ir
3
6.
2--
Upper lUffi.
101
developed from all data combined.
Data were also separated on the basis of whether saniples were
collected on the rising (n = 58) or falling (n = 74) limb of the hydrograph (Table 11, Figure 35).
Correlation coefficients for bedload and
CPOM equations were slightly greater for rising limb data than for falling limb data and the regression coefficients were slightly larger,
but the general form of equations fot both rising and falling limbs
was similar.
In contrast, regression coefficients for the lower rif-
fle equations were markedly larger than those for the upper riffle
equations.
These findings suggest that bedload transport may be more closely
related to stream discharge at the lower riffle, and that slightly
greater flows are required to produce the same rate of transport.
This
would be expected if, as previously stated, the pool tends to undergo
aggradation between storm seasons and results in a store of material
available for trax.sport during only moderately high streanif low at the
upper riffle.
Significantly higher streamflow is required to traxis-
port this material out of the pool and past the lower riffle.
The
lack of a difference between the equations developed for the data
separated on the basis of the limb of the hydrograph again indicates
that little hysteresis occurred during these storms.
Correlation coefficients for D50 ax.d D90 regression equations
were consistently low, so there appears to be a poor relationship between stream discharge and particle size in transport (Tables 9-11).
Except for the relationships developed for the upper riffle, D50 appears to increase with increasing discharge.
D90 also shows a slight
102
TABLE 11.
Relationships Between Bedload (BLD) Transport (BLD > 0.25
and BLD > 2.0
in kg.hr1), Coarse Particulate Organic
Matter (CPOM in kg.hri), Two Particle Diameters (D50 and
D90 in mm) and Stre.f low for the Rising (ii = 58) and Palling (n
74) Limbs of Storm Hydrograplis, Trap Bay Creek,
Chichagof Island, Alaska, During the Fall, 1980, Storm
Season
rising limb
ti ,2
Equation
r
2
F
1
a
b
(BLD > 0.25 mm) - 1858 Q165
0.70
**
**
*
) = 1035 Q175
0.67
*
**
*
119 Q1'84
0.79
**
**
*
(BLD > 2.0
(CPOM)
- 2.52
CD50)
(D90)
Q°14
0.03
13.3 Q°16
0.03
falling limb
1355 Q161
0.69
**
**
*
788 Q171
0.67
*
*
*
74 Q191
0.70
*
*
*
(D50)
2.56 Q°17
0006
(D90)
14.1 Q°19
0.10
(BLD > 0.25 u)
(BLD > 0.2
(CPOM)
mm) =
*
indicates significance at the 90% level of probability,
** indicates significance at the 95% level of probability.
2significance indicates that the regression coefficient differs
from zero.
ti
0
3
4
6
2
0.2
Q
0.67
15
2
(cr4)
73.96 Q191
788.3 Q1°
0.67
1.0
w> 2,0 mm)
0.2 0.4 0,6
-
QdI m)a-1-2
0.1
r - 0.69
135.2 Q'°l
(BLD >' 02 mm)
0.04 0.06
H
3, a
I-
4
2-
3-'H
4-
6
2-.
3-,-H
1.
6.
falilu! LIth
Bedload (BLD) and coarse particulate organic matter (CPON) vs. streamflow
(Q) relationships for rising limb (n = 58) and falling limb (n = 74) data
collected from Trap Bay Creek, Chichagof Island, Alaska
0.4 0.6 1.0
(cr014) 119.0
0.79
l04.9
(BLD ) 2.
1857.8 Q'.65
0.10
Q, tii3a'kir2
0.1
r2
(BLU > 0.25 a)
0.02 0.0( 0.06
Figure 35.
2-
4.
HiIng Lhb
104
increase with Increasing discharge, but none of the relationships is
very strong.
The data were also separated on the basis of storm event and relationships for Individual events were developed (Table 12, Figures
36 and 37).
Although there is an overall trend for bedload transport
to increase with increasing discharge evident in Figure 38, there is
considerable variation betweet the relationships for different storms.
In fact, the relationships developed for the 23 September storm were
so far below the scale of those for the other stores, that they could
not be included on the graphs.
There are no equations for the 18 Oc-
tober event because there were only two data points.
Regression coefficients for the same parameter vary by several
orders of magnitude for different storms (Table 12).
considerable variation among the test statistics.
There is also
The r
generally decreased when the data were stratified.
2
and F-values
Some of this de-
crease may be attributed to the smaller rnmiber of data points used
to develop the equations, but it also indicates that bedload trans-
port is a highly unsteady process that is likely to be influenced by
a number of factors in a relatively complex iarmer.
No single equation
examined here adequately characterizes bedload transport.
The realtionships obtained for D50 and D90 exhibit no definite
trend, either increasing or decreasing.
F-values for these equations
were very rarely significant,indicating a poor fit of the power equation.
"t-values" were occasionally significant for the "a" coefficient
(y-intercept).
This may mean that there is some "threshold
required to set a given particle size in motion.
discharge
However, the vari-
ability in the "a" coefficients indicates that there is more involved
105
TABLE 12.
Relationships Between Bedload Transport (BLD > 0.25 = and
BLD > 2.0
in kg.hr1), Coarse Particulate Organic Matter
(CPOM inkghr1), and Two Particle Diameters (D50 and Dg0
in ui) and Streamfiow (Q) for Individual Storm Events Which
Occurred During the Fall, 1980, Storm Season at Trap Bay
Creek, Chichagof Island, Alaska
r
2
F
1
ti2
a
b
23 September (n = 6)
(BLD > 0.25 =)
(BLD > 2.0
(CPOM)
(D50)
(D90)
5.7*1017Q115
6.0*106Q111
6,6*10 Q58
= 1.3*10 Q32
) =
-
8.5*1012Q8.2
24 September (n
10)
(BLD > 0.25
5440 Q146
491 Q291
864 Q_15
2.2
)
(BLD > 2.0 =)
(CP0M)
(D50)
(D90)
=
=
- 0.07
- 1.54 Q
28 September (n
(BLD > 0.25
(BLD > 2.0
)
(CPOM)
(D50)
(D90)
0.29
0.16
0.69
0.55
0.37
**
*
**
**
**
**
**
**
**
**
10)
0.48
)
0.14
0.08
0.38
0.02
0.13
- 185.1 Q022
- 60.3 Q174
109.3
-
0.84 Q07
3.16 Q
0.07
0.01
0.58
0.12
0.18
**
**
**
**
**
**
**
**
**
**
**
**
**
**
30 Sept. - 1 Oct. (n = 14)
(BLD > 0.25
(BLD > 2.0
(CPOM)
(D50)
(D90)
1 October (n
)
- 1950
) = 1476 Q193
= 67.3 Q048
= 4.47 Q
= 9.84 Q
9
**
32)
(BLD > 0.25 =) = 1632 1.53
(BLD > 2.0
) = 1075 Q154
(CPOM)
=
101 Q048
(D50)
= 3.26 Q
(D90)
0.57
0.45
0.53
0.02
0.06
= 15.1
Q°21
0.59
0.58
0.60
0.26
0.13
**
**
*
**
**
**
**
**
**
**
**
**
*
-
106
TABLE 12. -
Continued
t1 ,2
r
n) - 42254 3037
(BLD > 025
(BLD > 2.0
14368
)
(cPOM)
0.74
2.09 Q
64.5
-
(Pso)
(D90)
5 October (n
(BLD > 0.25
(BLD > 2.0
1
a
b
061
**
**
**
0 50
0.26
0.01
0.12
**
*
**
**
0.35
0.43
0.58
0 27
0 21
**
**
**
20)
10588
13183
1230
m)
m)
(CPOM)
2.99
- 858
(D50)
(D90)
= 319 Q
**
**
**
**
**
**
**
**
**
**
0.69
0.61
0.16
0 09
0 81
**
**
**
**
*
**
**
*
**
**
**
0.40
**
**
**
**
**
**
**
**
**
**
**
12)
(11
(BLD > 0.25
(BLD > 2.0
F
14)
2 October (n
7 October
2
966050
163305
2285
m)
m)
(CPOM)
(D50)
(D90)
7,4
-
5. 7.
o
91
0001
16 October (n - 14)
(BLD > 025 m)
(BLD > 20
(CPOM)
717.8
t) - 403e5
2L0
1q61
041
0006
=
=
=
CD50)
(D90)
2q41
8.73 Q
O.6
0
0.09
**
indicates significance at the 0.90% level of probability;
** indicates sigtiificance at the 0.95Z level of probability.
2significance indicates that the coefficient differs from zero.
NOTE:
The 18 October event was not evaluated because there were
only two data points.
2
.MH 3
H
f.4
0
6
NOTE:
Figure 36.
0.2
30 Sept.
(1.9
7 OiL.
0.4 0.6
20 Sept.
16
5 Oat.
1
t.
0
1 0t.
H
H
I
2.
1
2
-
2-
L
28
I
0.04 0.06
0.1
///7Oat.
Sept.
16
Oat.
a
0.2 0.4 06
24 Sept. j'30 Sept.
2 Oat.
lU)> 2.0
i;o
Oat.
3ir'ksr
Relationships are depicted only over Ihe range of streamfiow
observed. The 23 Sept. and 18 Oct. events are not depicted.
(see text).
Plotsalso inclucje some relationships that were
not statistically significant (see Table 12).
Q
q, 3,-1-2
Bedload (BLD) vs.
treamf low (Q) relationships for individual storm
events for data colected from Trap Bay Creek, Chichagof Island,
Alaska, during the Fall, 1980, storm season
0.02 0.04 0.06 0.1
24 Sept.
bill > U.'5
108
Figure 37.
Coarse particulate organic
matter (CPOM) vs. streamflow
(Q) relationships for individual storm evexlts for data
collected from Trap Bay Creek,
C1iichagof Island, Alaska, during the Fall, 1980, storm
season
NOTE:
Relationships are depicted only
over the range of streainflow
observed0 The 23 Sept. and 18
Oct. events are not depicted see ext Plots also include some reltIonships that
were no statistically significant (see Table 12).
109
in determining what particle size will be set in motion than discharge
alone.
The data were also examined to see if stratification by sampling
site, hydrograph limb, and site and hydrograph limb combined for each
storm to see whether correlations between streamflow and the parameters
being exiined would improve.
3.
These equations are included in Appendix
In general, test statistics for relationships for the two size
fractions of bedload and CPOM improved when the data were stratified
by spling site for each storm.
0.25
Total bedload transport (BLD >
). was highly correlated with streamf low during the 1 October
event at the lower riffle Cr2 - 0.96), but very poorly correlated with
streamflow at the upper riffle during this storm Cr2 = 0.06).
BLD >
2.0 =, CPOM, D50 and.090 were also fairly well correlated with discharge during the 1 October storm at the lower riffle.
Bedload trans-
port remained fairly well correlated with streamflow during the two
events following the 1 October storm, 2 October and 5 October.
The
relationships developed for all other events at the lower riffle had
relatively low correlation coefficients.
In contrast, correlation coefficients for relationships developed
for the upper riffle were highest for the three storms preceding the
1 October event, 24 September, 28 September, and 30 Sept.-1 Oct., and
were relatively high for the 7 and 16 October events.
Possibly, bed-
load transport and streaflow are more strongly related during lesser
agriitude events at the upper riffle than at the lower riffle because
of supply limitations.
Material may be available for transport past
the upper riffle during low magnitude events, but it is deposited in
110
the pool and is not readily available for transport past the lower
riffle.
Lower magnitude events may thus result in a build-up of
material in the pool but only transport material past the lower riffle sporadically, so there is not a good realtionship between dis-
charge and bedload transport at the lower riffle during these events.
Material transported to depositional areas during lower magnitude events may than become available for transport past the lower
riffle only during the following events of sufficient magnitude, such
as the 1 October storm at Trap Bay.
This storm may have initiated
transport of material out of the pool and made the remaining stored
material available for transport during the following two events.
charge could th
Dis
become the dominant factor in bedload transport be-
cause the lower riffle was no longer supply limited.
The poor relationship between discharge and bedload transport at
the upper riffle during the 1 October event may indicate that there
was little sediment available for transport.
The estimated amount
of sediment transport past the upper riffle was much less than that
for the lower riffle during this storm, and estimated sediment transport at the lower r.iffle only exceeded that at the tipper riffle during
the 1 and 7 October events (Table 8).
A lack of available sediment
for transport past the upper riffle during the 1 October storm could
have resulted in both a poor relationship between transport and
streamf low, and in a low amount of transport.
CPON relationships did not show the same trends as did the bedload transport relationships.
In general, CPON was more strongly
related to discharge than bedload transport, possibly because CPON
111
supply is more directly related to streaaflow, while bedload supply
can be limited by the size and arrangnent of particles on the bed.
Relationships developed for the data when they were stratified
by rising vs. falling limb on a storm by storm basis are also included
in Appendix 3.
r2-values for these relationships were higher for
rising-limb relationships for all parameters than they were for falling-limb relationships.
That there appears to be a stronger relation
ship between bedload and CPOM transport and streamf low during rising
limbs than during falling limbs was not obvious when all the storms
were considered together.
This may be a result of the greater ni.ber
of data points for falling limbs (n - 74) than for rising limbs
(ri = 58) or it may be that there is less storm-to-storm variation in
the falling limb data.
Finally, data were stratified according to storm, sampling site,
arid hydrograph limb (Appendix 3).
The relationships were generally
similar to those developed using various composites of the data.
Test
statistics were significant for the relationships for bedload and CPOM
for the same events that had significant relationships when composites
of the data were used (24 September, 1 October, and 5 October).
There
is considerable variation among regression coefficients for different
storms and between rising and falling limbs of the same storm.
There
is also variation in regression coefficients for relationships developed for the same hydrograph limb between the upper and lower riffles.
Variations in relationships tended to be considerably greater
for falling limbs than for rising limbs.
Obviously, the natural spa-
tial and temporal variation in bedload transport probably precludes
112
the development of a single general equation which will apply during
all storms.
Correlation coefficients for equations developed for the larger
size fraction of inorgauic bedload (BLD > 2.0
) tended to be slightly
lower than those for relationships of BLD > 0.25
and streaf1ow.
Apparently, transport of larger particles (medium to coarse gravel)
is not as strongly related to discharge.
2.0
"a" coefficients for (BLD >
) relationships were generally about half as great as those for
(BLD > 0.25
.greater.
) relationships, while 11b" coefficients were generally
Thus, about half of the material transported at a given
discharge was composed of
ediu
to coarse sand and gravel, while the
remainder was composed of larger particles.
The mass of larger par-
ticles in transport tended to increase at a slightly greater rate
with increased discharge than did the total mass in transport.
CPOM relationships tended to have significant test statistics for
the same events for which bedload relationships were significant.
However, different processes appear to control the supply arid trans-
port of organic and inorganic particulate matter, even when both are
related to streaflow.
This is evident if regression coefficients
for bedload and CPOM relatioships for the same storm, sampling site,
and hydrograph limb are compared.
Organic material enters the system sporadically in the form of
coarse particulates (leaves
twigs, needles, and bark).
This material
is broken down to fine and very fine particulates by aquatic organisms
and by mechanical processes, and then becomes available for transport
in suspension.
CPOM must first be colonized by bacteria and fungi
113
before it becomes a good food source for aquatic organisms (Cummins,
1974).
Deciduous plant material is more readily utilized as a food
source by these organisms than is evergeen plant material; thus it
is reduced to fine particulate form more quickly (Cnins, 1974).
Therefore, several factors are important in determining the supply
and ultimate partitioning of the organic load of the stream, including the
ate and type of material entering the stream, the rate of
consption of organics by aquatic organisms, and the availability of
the material for transport (size and location).
Even during the fall,
when organic inputs are greatest, Trap Bay Creek appears to be supply-limited in organic material available for transport because so
little shows up in suspension.
Data collected using the Helley-Smith sampler indicated that during the 1980 Fall storm season at Trap Bay Creek, streamf low was only
one of the factors determining bedload transport, that there was a
general increase in material transported with increasing discharge,
and that transport appeared to occur in pulses or "waves."
Particle
size diameter did not appear to be related to discharge, nor did there
appear to be any relation between particle size diameter and the total
mass of material in transport.
Bedload Transport at Flynn Creek, Oregon vs Trap Bay, Alaska
Flynn Creek is a small second order stream in the Alsea River Basin of Western Oregon.
Sediment sampling facilities were installed in
1976; several researchers have conducted bedload sediment transport
114
studies there since 1976.
Both the Flynn Creek and Trap Bay Creek
Watersheds receive most of their precipitation in the form of light
to moderate intensity, long duration frontal storms, wIth 90% or more
occurring from Novber to May.
The total precipitation and average
tperature of the two areas are very similar.
In contrast to Trap Bay, however, soils at Flynn Creek are relatively deep and are derived from sandstone bedrock.
(Pseudotsuga
overstory.
Douglas fir
e.nziesii) and red alder (Ai.nus rubra) dominate the
The drainage area of Flynn Creek is 218 ha (O'Leary,
1980) while that of Trap Bay is 1355 ha0
The average elevation of
Flynn Creek is 320 m and the relief ratio is 0.13, while these parameters for Trap Bay Creek are 590 m and 0.24, respectively.
Flynn
Creek is about one-third as long, one-tenth as wide, and one-ninth
as deep as Trap Bay Creek.
Thus, there are several differences be-
tween these two streams which must be kept in mind when comparisons of
sediment transport are made.
Table 13 lists bedload and CPOM transport relationships developed
for Flynn Creek for the 1978 (O'Leary, 1980) and 1979 (Edwards, 1979)
water years.
These relationships were developed using the same power
equation that was used in developing relationships for Trap Bay Creek.
The data presented by O'Leary (1980) were collected with a vortex sampler.
Data presented by Edwards (1979) were collected with a vortex
sampler at the fishtrap and with a Helley-Smith sampler with a large
collection bag at the riffle site.
It is at once obvious that the realtionships developed from data
collected at the riffle site at Flynn Creek are more similar to those
115
TABLE 13.
Bedload (BLD) and Coarse Particulate Organic Matter (CPOM)
Relationships for Flymi Creek, Oregon, and Trap Bay Creek,
Chichagof Island, Alaska
Site of
collection
Flynn Creek,
Fish Trap
Date
11/24-25
1978
12/ 2-3
1978
12/13-15
- 1978
Flynn Creek,
Riffle Site
2/7
[1979
1979 WY
2/ 6-8
Flynn Creek,
Fish Trap
Trap Bay Cr.,
data from
both riffles
-
2/7
1979
Equation
BLD > 0.25
=
BLD > 0.25
= 110
98 Q824
r2
Source
0.66
0' Leary,
1980
Q141
BLD > 0.25
2.46
2 64
=1536
BLD > 2.0 mm = 770 Q
1.98
BLD > 0.25 mm
137
2 80
CPOM
= 571 Q
BLD > 0.25
10/1
1980
BLD > 0.25
BLD > 2.0
CPOM
represents streaflow in
131 Q
= 250
BLD > 0.25
BLD > 2.0
CPOM
4.13
5 27
i = 766
BLD > 2.0 =
CPOM
0.32
694 Q543 0.55
BLD > 025
1979 WY
1980
storm
season
1
Q255
1.53
i =1632 Q173
=1075 Q1
101 Q
=1534
m3sici2
877 Q186
89 Q
'
0.79
0.71
Edwards,
1979
0.90
0.90
0.93
0.92
0.96
0.59
0.58
0.60
0.69
0.66
0.77
Edwards,
1979
116
developed for Trap Bay Creek than are those developed from .data collected at the Fish Trap.
There are two reasons why this is to be ex-
pected: (1) A vortex sampler was used at the Fish Trap while a HelleySmith sampler was used at the riffle site, (2) Sand-sized particles
are the dominant particles in transport at the Fish Trap while gravels tend to to dominant at the riffle (Edwards, 1980).
In general, "b" coefficients for the equations for Flynn Creek
are significantly greater than those for relationships for Trap Bay
Creek, indicating that transport tends to increase more rapidly with
discharge at Flynn Creek.
The "as' coefficients for Flynn Creek re-
lationships tend to be somewhat less than those for Trap Bay Creek.
This probably reflects the difference in the average particle sizes
of bed material in these two streams, with that of Trap Bay Creek
being somewhat larger and more angular (personal observation).
Relationships developed for the peak flow event of 1979 at Flynn
Creek, 7 February, and the peak flow event of 1980 at Trap Bay Creek,
1 October, are similar for the riffle site at Flynn Creek.
Edwards
(1980) reported an estimate of inorganic sediment yield of 2.6 tonnes
and sri organic sediment yield of 1.5 tonnes at the Fish Trap for the
24-hour period on 7 February 1979, the annual peak flow event.
This
storm had a peak discharge of 0.75 m3skm2. For the riffle site
at Flynn Creek for the same time period, inorganic sediment yield was
estimated to be 13.0 tomies and organic sediment yield was estimated
to be 1.7 tonnes.
For Trap Bay Creek during the 1 October event which
3-1 1i -2
had a peak flow of 2.56 m s
, based on the relationships develop-
ed for this storm (Table 12) and using data from both riffles, total
117
inorganic sediment yield was 8.0 tonnes and total organic sediment
yield was 0.5 tonnes.
This is
not what the difference in the size and
discharge between the two streams would lead one to expect.
The low
total sediment yield for Trap Bay Creek probably again reflects the
difference in the size and shape of bedload particles between the two
streams
The major factor in influencing the size and shape differences
in bedload sediment is the geology of the watershed.
The sandstone
bedrock underlying the Plyn Creek watershed is relatively easily
weathered and breaks down to form an abundance of sand-sized particles
(O'Leary, 1979).
Thus, most of the bedload sediment yield consists of
sand-sized particles (Edwards, 1980).
Granitic parent material sup-
plies some of the material transported as bedload by Trap Bay Creek.
This material is relatively resistant and, when broken down, the particles may compose a large proportion of the armor layer and tend to
orient themselves to flow in such a way as to resist transport.
Thus,
a greater discharge is necessary to initiate bedload transport at
Trap Bay Creek than at Flynn Creek because of the larger size and the
orientation of the armor layer particles.
The low total sediment
yield at Trap Bay Creek probably reflects the fact that more nergy is
required to dislodge and maintain transport of these larger, angular
particles.
118
VII.
CONCLUSIONS
Based on general observations and a planimetric survey of the low-
er 1890 m of the strea9 the channel morphology of Trap Bay Creek is
largely influenced by inputs of large organic debris in conjunction
with flows ad bedload transport.
Tree-sized material serves to trap
small debris and sediment, can result in localized deposition and
scour, contributes to batik stability, and serves as cover for salmonid s.
Because of the lithology of the watershed, during average storm
events, more sediment is transported out of the system in the fort of
bedload than is transported as suspended sediment.
Both suspended
and bedload transport appear to be supply limited, however, so that
any increased availability in sediment is likely to result in increased transport rates.
Although suspended sediment transport appears to be normally much
less than bedload transport, any future research on suspended sediment
transport in this stream in Trap Bay Creek, should include modification of analytical procedures and equipnLent to alleviate problems en
countered in this study.
These modifications might include the ad-
dition of an accurate weighing scale (for filter discs) and a longrunning power source (for turbidimeter warm-up) at the field installation.
4
Bedload transport is a dynamic process which is generally rela-
ted to stream.flow.
Based on sampling conducted during Fall 1980 at
a pool-riffle sequence on Trap Bay Creek, events with a magnitude of
less than one year return period appeared mainly to transport material
119
to the pool where it was temporarily stored.
Transport through the
pool-riffle sequence occurred when the magnitude of the storm event
was great enough to transport material out of the pool and past the
riffle.
Transport of material dislodged by greater magnitude events
may continue during successive low to moderate storms until some form
of armoring takes place in the pool or some other form of supply limitation begins to dominate the transport process.
Rating curve relationships indicated a general increase in bedload
transport with Increased discharge, but relationships developed for
individual events show considerable variability.
Rating curve rela-
tionships developed for this study period appear to be site and storm
specific for Trap Bay Creek.
Over the range of streamflows dur1n which sampling took place,
particle size diameter did not appear to be related to discharge, nor
did it appear to be related to the total amount of material in transport.
Bedload transport during the study period appeared to be more
strongly controlled by stream discharge on the rising limb of the hydrograph than on the falling limb.
Supply limitations or partial
rearmoring of the streambed may have influenced transport on the falling limb.
CPOM transport is generally more strongly related to dis-
charge than is bedload transport.
The difference in the geology and lithology of Trap Bay and Flynn
Creek, as well as watershed and stream size differences, apparently
resultsin some interesting contrasts in sediment transport between
these two systems.
Suspended sediment is a much larger portion of
120
total sedimetit transport in Flynn Creek, atid smaller particle sizes
characterize the bedload portion relative to bedload in Trap Bay
Extreme spatial and tporal variability iii bedload transport
is characteristic of both streams Total sedimetit yield, however,
Creek.
appears to be greater for Flynn Creek ad transport appears to increase more rapidly here with increasitig streamflow than in Trap Bay
Creek.
This may reflect the differences in the size and shape of
bedload particles between the two streams, which are ultimately a restilt of the geologic differetice between the two watersheds.
100
The hatid-held Helley-Smith sampler provided a means of collecting
bedload transport data iii a relatively inaccessible area.
Samples
obtaitied using the He1ley-Sith do not incorporate all of the tenpor-
al and spatial variation in bedload tratisport, but they do appear to
provide a useful index to relative transport rates.
121
VIII.
LITERATURE CITED
Adams, J.N.
1980. Variations in gravel bed composition of small
streams in the Oregon Coast Range. MS thesis, Oregon State
University, Corvallis. 159 pp.
American Public Health Association.
1976. Standard Methods for Examination of Water and Wastewater, 14th Edition. American Public Health Assoc., New York, N.Y.
1193 pp.
Bagnold, R.A.
1977. Bedload transport in natural rivers.
Resour. Res. 13(2):303-3j.2.
Water
Beschta, R.L.
1978. Long-term patterns of sediment production following road construction and logging in the Oregon Coast Range.
Water Resour. Res.
14(6):1011-1016.
1979. Hydrologic research in southeast Alaska - an overview
(unpublished report). Dept. of Forest Eng., Oregon State University, Corvallis. 21 pp.
1981. Increased bag size improves Helley-Smith bed load sampler for use in streams with high sand organic matter transport.
Pp. 17-25.
IN: Erosion and sediment transport measurements,
Proc. of the Florence Symp., June 1981. Inter. Assoc. Hyd.
Beschta, R.L. and W.L. Jackson. 1979. The intrusion of fine sediments
into a stable gravel bed. J. Fish Res. Board. Canada. 36(2):
204-210.
Brusven, M.A.
1980. Secondary productivity in aquatic ecosystems -the effects of laxtd use practices on invertebrates (unpublished
report). Dept. of Entomology, University of Idaho, Moscow.
20 pp.
Chapman, D.W.
1961. Factors determining production of coho salmon,
Oncorhyncus kisutch, in thee Oregon streams. Ph.D. dissertation,
Oregon State University, Corvallis.
214 pp.
Clayton, J.L.
1981. Magnitude and iuiplications of dissolved and sediment elemental transport from small watershed. Pp. 83-98.
IN:
D.M. Baumgartner, (ed.). Interior West Watershed Management.
Conmittee on Erosion and Sedimentation, AGtJ Hydrology Section.
1977.
Research needs in erosion and sedimentation.
Amer. Geophys. Union
Trans.
58(12):1076-1083.
Cumins, K.C.
1974. Structure and function of ecosystems.
24:631-641.
Bio-science
122
Curry, R.R.
1973. Geologic and hydrologic effects of even-age managesent on productivity of forest soils, particularly in the Douglasfir region.
Pp. 137-178. IN: R.K. Herman and D.P. Lavender,
(eds.). Even-age Management0
Drufel, L.W., LW. Enixiett, V.P. Schneider, and J.V. Skimier,
1976.
Laboratory calibration of a He1ley-Smith bedload sampler. USD1
Geol. Surv. Open File Report 76-752.
63 pp.
Dunne, T., and L.B. Leopold. 1978. Water in Environmental Plaing0
W.R. Freeman and Company, San Francisco. Pp. 591-708.
Edwards, R.E. 1979. Sediment transport and channel morphology in a
small streani in Western Oregon. MS thesis, Oregon State University, Corvallis
114 pp.
E.ett, W.W.
1975. The channel and waters of the Upper Salmon River
Alsea, Idaho. USD1 Geol. Surv. Prof. Paper 870-A. Pp. A85-A86.
1979. A field calibration of the Helley-Smith bedload sanipler
UADI Geol0 Surv. Open Pile Report 79-411, 96 pp.
_____
1981. Measurement of bedload in rivers. Pp. 3-15.
IN: Erosion
and Sediment Transport Measurients, Proc. of the Florence Symp.,
June, 1981. Inter. Assoc. Hydro. Sci. Publ. no. 133.
Graf. W,H.
1971. Hydraulics of Sediment Transport.
Co., New York, N.Y.
513 pp.
McGraw-Hill Book
Haddock, D.R.
1978. Modeling bedload transport in mountain streams of
the Colorado Front Range. MS thesis, Colorado State University9
Ft. Collins.
74 pp.
Hall, J.D, and J.T. Krygier.
1967
Progress report: studies on effects of watershed practices on streams. Report to the Fed. Water
Pol. Control Admin on Res. Grant WP-423. Oregon State University, Corvallis. 95 pp.
Harr, R.D.
1976. Forest practices and streamflow in Western Oregon.
USDA-Forest Ser. Gen. Tech. Report PNW-49. 18 pp.
Harris, A.S., O.K. Hutchinson, W.R. Meehan, D.N. Swanston, A.E. Helmers, J.C. Hendee, and T.M. Collins. 1974. 1. The setting -- the
forest ecosystems of southeast Alaska. USDA For. Ser. Gen. Tech
Report PNW-12. 40 pp.
Harris9 D.D0 and R.C. Williams.
1971. Streaiuf low, sediment-transport,
arid water temperature characteristics of three sma.11 watersheds
in the Alsea River Basin, Oregon. USD1 Geol. Surv. Circ. 642.
21 pp.
123
Hayward, J.A., and A. 3. Sutherland.
1974. The Torlesse Stream vortex-tube sediment trap. 3. Hydrology (New Zealand) 13(1):41-53.
Heede, B.H.
1975. Mountain watersheds and dynamic equilibrium.
Pp.
407-420.
IN: Proc. Watershed Mgmt. Symp., Amer. Soc. Civil Eng.
Irrig. and Drain. Div., August 11-13, 1975. Logan, Utah.
Helley, E.J., and W.W. SuLith.
1971. Development and calibration of .a
pressure-differential bedload sampler. USD1 Geol. Surv. Open
File Report, Menlo Park, Calif. 18 pp.
Hillel, D.
1980. Fundamentals of Soil Physics. Academic Press Inc.,
New York, New York. 413 pp.
Hubbell, D.W.
1964. Apparatus and techniques for measuring bedload.
USD1 Geol. Surv. Water-Supply Paper 1748.
U.S. Govt. Printing
Office, Washington, D.C. 73 pp.
Hynes, H.B.N.
1970. The Ecology of Running Waters.
University of Toronto Press, Toronto, Canada.
Pp. 203-245.
Inghram, Mark.
1979. Kadashan precipitation data compilation and
analysis (unpublished report).
USDA For. Ser., Tongass Nati.
For., Sitka, Alaska. 38 pp.
Johnson, C.W., R.L. Engleman, J.P. Smith, and C.L. Hanson. 1977.
Helley-Smith bedload samplers.
Amer. Soc. Civil Eng. 3. Hyd.
Div. 103(EY 10):1217-1221.
K.lingeman, P.C.
1971. Evaluation of bedload and total sediment yield
processes on small mountain streams. Pp. 58-169.
IN: J.T. Krygier (ed.). Studies on effects of watershed practices on streams.
Oregon State University, Corvallis.
Klingeman, P.C., and R.T. Milhous.
1970. Oak Creek vortex sampler.
IN: Proc. of the 1971 Hyd. Div. Specialty Conf., Amer. Soc. Civii Eng. August 18-20, 1970. University of Iowa, Iowa City.
-
Langbein, W.B., and L.B. Leopold.
1968. River channel bars and dunes theory of kinematic waves. USD1 Geol. Sur. Prof. Paper, 422L.
Leopold, L.B., M.G. Woman, and J.P. Miller. 1964. Fluvial Processes
in Geomorphology. W.H. Freeman and Company, San Francisco.
522 pp.
Leopold, L.B., and W.W. Emmett.
1981. Some observations on the movement of cobbles on a streambed. Pp. 49-59.
IN:
Erosion and
sediment transport measurement, Proc. of the Florence Symp.,
June, 1981. mt. Assoc. Hydrol. Sd. Pubi. no. 133.
124
1978. Morphometric indices of stream.flow and sediment
yields frommonntauwatersheds in Western Oregon. USDA For.
Ser. Project Report P.O. 00-04 TO-9-73. Siuslaw Natl. Forest,
Corvallis, Oregon. 82 pp.
Marston, R.A.
1975. Sedimentation in relation to logging activities
in the mowitains of central Idaho. Pp. 74-82. Reprint from
"Present and prospective technology for predictitg sediment
Megahan, W.F.
yields ad sources."
28-30, 1982.
USDA Sed. Lab0
Oxford, Nissippi. Nov.
Milhous, R.T., and P,C. Klingematt. 1973. Sediment transport system in
a gravel-bottomed stream. IN: Hydraulic Engineering and the En-
viroment, Proc. of the 21st Annual Hyd. Div. Spec, Conf., Amer.
Soc. Civil Eng., August 15-17, Montana State University, Boznan
Molnau, et al. 1975. Sediment transport estimation in the Central
Idaho Batholith. Pp. 377-392. IN: Watershed management, Amer.
Soc. Civil Eng. Symp. Proc. Irr. and Drain0 Div., August 11-13,
1975. Logan, Utah. Publ. ASCE, New York, NY.
Neter John and Wiili
Wasserzan.
1974
Applied Linear Statistical
Models. Richard D. Irwin, Inc. Homewood, Illinois.
842 pp.
Nordin, C.F. 1981. The sediment discharge of rivers - a review.
Pp. 3-47. IN: Erosion and sediment transport measurements,
Proc. of the Florence Syrnp., June, 1981. Inter. Assoc. Hydrol.
Sci. Publ. no. 133.
Nuttal P.M. 1972. The effect of sand deposition upon the macroinvertebrate fauna of the River Camel, Cornwall. Freshwater Biol.
2: 181-186.
1980. Bedload transport in an Oregon Coast stream.
MS thesis, Oregon State University, Corvallis. 107 pp.
O'Leary, S.J0
Park, C.C. 1976. Man-induced changes in stream channel capacity.
Pp. 121-144. IN: K.J. Gregory (ed.). River Channel Changes.
John Wiley and Sons, New York, N.Y.
Phillips, R.W. 1971. Effects of sediment on the gravel environment
and fish production. Pp. 64-74. IN: Forest land uses and streaxa
environment, Proc. Syxnp., 1970, Oregon State University, Corval-
lis
1978. Prediction techniques for potential changes in
sediment discharge due to silvicultural activities. Unpublished
paper presented at Amer.. Soc. Civil Eng. Natl. Metting, April,
1978, Pittsburgh, Pa. 26 pp.
Rosgen, D.L.
125
Satterlaund, D.
1972. Wildland Watershed Management.
Co., New York, N.Y.
174 pp.
Ronald Press
Scbmeige, D.C., A.E. Helmers, and D.M. Bishop.
1974. 8. Water -- the
forest ecosystem of southeast Alaska. USDA For. Ser. Gen. Tech.
Report PNW-28.
26 pp.
Sidle, R.C., and D.N. Swanston. 1982. Analysis of a small debris slide
in coastal Alaska. Can. Geotech. 3., 19, 040-000:167-174.
Simons and Senturk.
1976. Sediment Transport Technology.
508.
Water Resour. Pubi., Fort Collins, Colorado.
Pp. 504-
Swanson, F.J.
1980. Geomorphology and ecosystems. Pp. 159-170.
IN: Richard W. Waring (ed.). Forests: fresh perspectives from
ecosystems analysis, Proc. 40th Annual Biol. Coll., 1979.
Swanson, P.3., and G.W. Lienkaemper.
1978. Physical consequences of
large organic debris in Pacific Northwest streams. USDA For.
Ser. Gen. Tech. Report PNW-69.
12 pp.
Swanson, P. 3., R.L. Predriksen, and F.M. McCorison. (In press).
Material transfer in a Western Oregon forested watershed.
IN:
R.L. Edmonds (ed.). The Natural Behavior and Response to Stress
of Western Coniferous Forests. Dowden, Hutchinson, and Ross.
Publishing Co., Stroudsburg, Pa.
Swanston, D.N.
1974. 5. Soil mass movement - the forest ecosystem
of southeast Alaska. USDA For. Ser. Gen. Tech. Report PNW-83.
15 pp.
Swanston, D.N., and P.3. Swanson.
1976. Timber harvesting, mass
erosion, and steepland geomorphology in the Pacific Northwest.
Pp. 199-211.
IN: D.R. Coates (ed.). Geomorphology and Engineering.
Dowden, Hutchinson, and Ross Publishing Co., Stroudsburg, Pa.
Swanston, D.N., and W.J. Walkotten.
1974. Tree rooting and soil stability in coastal forests of southeastern Alaska. Study No.
FS-NOR-1604:26, PNW Range and Exp. Sta., Forest Sciences Lab.,
Juneau, Alaska.
Vanoni, V.A. (ed.).
1975. Sedimentation Engineering.
Civil Eng. Publ., New York, N.Y.
745 pp.
Amer. Soc.
Viereck, Leslie A. and E.L. Little, Jr. 1972. Alaska Trees and Shrubs.
Agric. Handbook No. 410. USDA Forest Service, U.S. Govt. Printing Office, Washington, D.C.
265 pp.
Water Resources Atlas for Alaska.
Region X, Juneau, Alaska.
April, 1979.
USDA For. Serv,
APPENDICES
A.PPEI1DIX 1.
List of Conon ad Scientific Names of Plants
and Animals Referred to in This Paper
126
Plants
Sitka spruce
Western hemlock
,
. Picea sitchensis (Bong.). Carr.
Tsuga heterophylla (Raf0) Sarg.
.
Red alder
Alnus rubra Bong.
Lodgepole pine
Pinus contorta Dougl.
Western redcedar
.
Thuja plicata Doun.
Blueberry
Vaccinium alaskaense Howell.
Huckleberry
V. parvifolium Sm.
Salmonberry
Rubus pectabilis Pursh.
Devils club
Oplopanax horridus (Sm.) Miq.
Skunk cabbage
Lysichitum americanum Hult. and St. John
Sedge
Carex L.
Nettles
TJrtica Lyallii S. Wats.
Animals
Beaver
Castor canadensis L.
Pink salmon ....... .
Oncorhynchus gorbuscha Walbaum0
Coho salmon
O
Dolly varden char
Salvelinus malma Walbaum.
kisutch Walbau.
APPENDIX 2.
Equations
Flows for
Storms of
Chichagof
For Predicting Mean Annual Flow, Mean Monthly
August Through November, and Peak Flows for
Various Return Periods for Trap Bay Creek,
Island, Alaska, Taken from the Water Resources
Atlas for Alaska (1979)
127
1
113 103
A
- .0312 P
= 29.45 cfs
an
90% C.I. - 25-34
Mean annual flow,
Mean August flow,
A
0.97
.0013 Pm 1.43 A952 V°181E671C'179= 36.7 cfs
= 0.96 90% C.I. = 18-42
mean monthly precip.
P
7.6 in.
44.35 cfs
P107 A99 E'34
2154
Mean September
.0564
flow,
R2 = 0.97 90% C.L
P
114 in.
m
Mean October
Qo
flow,
- 1.26 m981 A1°5 C169 - 38.8 cfs
R2 = 0.97 90Z C.I.
14.25 in.
P
25-57
m
- 4.03 m838 A1°5 T°°18
Mean November
flow,
R2 - 0.97 90% C.I.
10.45 in.
2-year peak flow, Q
2
5-year peak flow, Q
= 124
L24
17.8
902
25-year peak flow,
19.8
E
=S2Ocfs
375=910
0.461
E
P12° A907 -.346
L
703cfs
470-1080
P115 A898 L352
R2 - 0.93 90% C.I.
23.7
-.477
-V.337
R2 = 0.93 90Z C.I.
10-year peak flow, Q
10
22.0 cfs
14-27
L
A
R2 = 094 90% C.I.
P
362=
C
E
-.417
=84Ocfs
535-1145
P12 A905 L355 E408 =946 cfs
0.93 90Z C.I.
595-1490
-.356
371
1079cfs
50-year peak flow, Q
50
26.2 P1"'09 A903 L
100-year peak
30.3 P1°6 A904
E°371 = 1196 cfs
- 0.91 90% C.I. = 710-1890
flow,
R
2
= 092 90Z C,I.
E
640-1650
P = mean axnual precipitation from map - 95 in.
A - basin area = 5.23 sq.mi.
L - proportion of basin in main channel lakes - used 1% to
avoid negative logarithni.
E = mean basin elevation:
E
+ E
+ .324(E.
E
1254 ft.
in
aye
in )
4 of basin aDove treeline = 3.74
T
C
south distance to Gulf of Alaska = 277.5 mi.
.
APPENDIX 3.
Relationships Between Bedload Transport, Coarse Particulate
Organic Matter Transport, Two Particle Diameters, and Streamflow Which Were Not Included in the Text
128
TABLE 16.
Relationships1 Between Bedload Transport (BLD > 0.25 =
and BLD > 2.0 nm), Coarse Particulate Organic Matter
Transport (CPOM), Two Particle Diameters (D50 and D90),
and Streamf low2 for Individual Storm Events at the Upper
Riffle, Trap Bay Creek, Chichagof Island, Alaska, During
the Fall, 1980, Storm Season
Storm
n
equation
r2
BLD > 0.25 =, kg.hr
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
5
5
6
16
7
10
6
7
3.10
884505 Q252
23742 Q127
1316 Q193
1846 Q03
966 A426
119058 Q204
28128Q80
2,5*10
927 Q
BLD > 2.0
23
24
28
30
Sept.
Sept.
Sept.
Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
5
5
6
16
7
10
6
7
-1
83
0.03
0.82
0.44
0.76
0.06
0.84
0.20
0051
'
,
kg.hr1
7*106
1302 Q1049
1142 Q2'1
1464 Q
567 Q.,
12015
48439Q1
8
2.7*10
516 Q '
0.06
0.70
0.54
0.64
0.07
0.48
0.36
0,70
0.59
CPOM, kg.hr'4
23 Sept.
24 Sept.
28 Sept.
30 Setp.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
5
5
6
16
7
10
6
7
s*io14
835 Q147
59 Q061
71 Q032
80
8
1.6 Q
2399 Q
0.1
25 Q
0
0.07
0.81
0.60
0.61
0.08
0.14
0.61
0.00
0.28
F3
a
129
TABLE 16. - continued
Storm
n
equation
r
2
F
3
t
a
b
D50,
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
12
2.7*10
3
5
8.12
Q21
0.02 Q020
3.94 Q039
5
4.81 Q02
2.50 Q13
6
16
0.36 Q216
7
10
16.93
4.10
2272.20 Qo 22
2.43 Q
6
7
0.45
0.83
0.04
0.06
0.07
0.17
0.52
0.21
0.06
**
**
**
*
**
**
**
**
**
D90,
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
5
5
6
16
2.9*1012
7
14.41
1.00
9.27
8.60
7.77
-o.2
3.21 Q043
o.i9
10
17.18 Q425
6
7
17093.10
0 4
7.77 Q
0.50
0.81
0.05
0.02
0.06
0.01
0.03
0.21
-.09
**
**
**
**
**
**
**
1No equations were developed for events represented by less than
three data points.
2
.
3-1
Streamf low, Q, is in m s
-2
kin
indicates significance at the 90% level of probability;
** indicates significance at the 95% level of probahility.
130
TABLE 17.
axid
Relationships1 Between Bedload Transport (BLD > 0.25
BLD > 2.0 nun), Coarse Particulate Organic Matter Transport
(CPOM), Two Particle Diameters (D50 and D90), and Streamflow2 for Individual Storm Events at the Lower Riffle, Trap
Bay Creek, Chichagof Island, Alaska, During the Fall, 1980
Storm Season
Storm
U
equation
r2
, kg.hr
2.5*106 3.95
1288
BLD > 0.25
23 Sept.
24 Sept.
28 Sept.
30 Setp.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
5
6
3715
2583 Q349
7
26384
4.o3
42970
0.09
i 26
504 Q
10
6
7
BLD > 2.0
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
F6 Oct0
l6Oct.
kg.hr
074
061
023
0.03 Q131
0.10
214 Qo8
030
16
7
10
6
7
3
5
5
6
16
7
10
6
7
Q241
Q212
Q406
Q468
Q419
0.01
i 31
290 Q
b
**
**
*
**
**
**
**
**
**
**
*
1
3
4
1503
1720
27326
37775
a
*
**
**
**
0.32
5
5
6
CPOM,
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
0.14
0.52
0.13
0.21
0.96
p2.57
i.90
16
F3
1
**
*
**
**
**
**
**
**
0.14
011
0.95
0.70
054
012
**
**
**
**
0.25
kghr1
13408
3.20
3.i4
1025 Q197
189 Q159
35 Q162
127 Q32
0.3 Q375
1057 Q48
0.0002
18Q o
04
0.47
0.71
0.59
0.14
0.65
0.43
0.72
0.03
0
**
**
**
**
**
*
**
**
**
**
**
*
**
*
131
TABLE 17. - continued
Storm
n
equation
r
2
a
b
D50,
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7
Oct.
16 Oct.
3
-9.9
5.6*1015
Q_14
Q_02
Q059
3.90 Q152
0.31
0.19
1.77
5
5
6
16
7
1202 Q097
10
6
4.09 Q04
0.64
2.49
7
o 03
Q
0.58
0.23
0.37
0
0.36
0.12
0.10
0.01
**
**
**
*
0.
D90, =
23 Sept.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
3
200.90
2.72 Q12
5
5
6
16
7
10
6
1.05
Q07
1.17
57.27
16.29
179.34
o.8o
Q032
2.o7
0
0.11
0.31
0.06
0.27
Q_oo3 0.40
0.03 Q181
0.13
0.06
164.06 Q0
11.05
7
Q
0.08
*
**
**
**
**
**
*
**
1No equations were developed for events represented by less than
three data points.
2
3-1 -2
Streamflow, Q, is in m s
indicates significance at the 90% level of probability;
** indicates significance at the 95% level of probability.
132
TABLE 18.
Relationships1 Between Bedload Transport (BLD > O25
and
BLD > 2.0 mm), Coarse Particulate Organic Matter Transport
(CPOM), Two Particle Diameters (D50 and D90), and Streamflow2 For the Rising Linibs of Storm Hydrographs For mdividual Storm Events at Trap Bay Creek, Chichagof Island,
Alaska, During the Fall, 1980, Storm Season
Storm
n
equation
r
2
F3
a
b
**
**
*
BLD > 0.25 mm, kg.hr
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
8345
185
5
10
187
6
14
1689
128345
2 38
4677 Q
9
12
BLD
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
10
6
14
9
12
2.51
2.0
0.58
0.07
0.15
0.74
0.81
0.45
*
**
**
**
**
**
**
**
**
**
, kg.hr
760
60 Q116
102
1095 Q42
52119 Q31
4764 Q
0.36
0.01
0.06
0.72
0.78
0.52
**
**
**
**
**
**
**
**
**
**
O82
**
**
**
**
**
**
**
**
**
**
**
**
0,2O
O.75
**
**
**
*
*
**
**
*
CPOM, kg.hr
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
10
6
14
9
12
3.41
2728 Q174
109 Q438
13772 Q220
107
1.4
2463 Q
4 70
O85
0.58
0.70
D50,
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
10
6
14
9
12
0.06
0.17
o 68
10.69 Q007
3.54
3.99 Q119
3.68 Q
0.68
0.12
0.02
0.01
0.03
0.20
133
TABLE 18. - continued
t3
Storm
n
equation
r2
p3
a
b
D90,
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
L13
10
3.16
0
6
48.14 Q03
14
1658 QQL4
9
12
.
81.66
24.60 Q0
0.76
0.18
0.17
025
0.21
O16
**
*
**
**
**
*
*
*
'No equations were developed for events represented by less than
three data points.
2
3-1 -2
Streamfiow, Q, is in m s
1
indicates significance at the 90% level of probability;
** indicates significance at the 95% level of probability.
134
TABLE 19.
Relationships1 Between Bedload Transport (BLD > 0.25 = and
BLD > 2.0 nun), Coarse Particulate Organic Matter Transport
(CPOM), Two Particle Diameters (D50 and D90), and Streamflow2 for the Falling Limbs of Storm Hydrographs For Indi
vidual Storm Events at Trap Bay Creek, Chichagof Island,
Alaska During the Fall, 1980, Storm Season
Storm
n
equation
BLD > 0.25
23 Sept.
24 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16
Oct0
, kg.hr
4
3948117 Q9
5
6
18
5
8
15
12
14
1424
1451
3337
222844
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
4
5
6
18
5
8
12
14
4
5
6
18
5
8
12
14
a
0.03
0.01
0.60
0.27
0.05
**
**
**
**
**
**
*
**
**
*
**
**
**
**
**
*
**
**
**
**
**
**
**
**
**
**
*
**
**
b
0.07
040
kg.hr1
,
992 Qj
1075
893
75617
166805
15066
F3
015
°
718 Q
0
0.. 03
82
Q105
Q518
Q633
Q388
Q165
0.47
024
0.23
0.14
0.08
0.41
404 Q
CPOM9
23 Sept.
24 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
Q086
Q175
Q602
Q322
11337 Q11
BLD > 2.0
23 Sept.
24 Sept.
30 Sept.- 1 Oct.
6
i
r2
-1
*
k.hr1
0.27
19896
3 Q209
61 Q035
78 Q20,
254 Q4
1216
0.0006 Q0 o
21 Q
°
0
.
0.77
0.14
0.96
0.13
0.01
0.06
D50, mm
23 Sept.
24 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
4
4,27 Q°1
0
5
0.22 Q06
6
4.74 Q051
18
3.13 Q99
5
8
45.29 Q083
12
14
51.59 Q9
0.09
0.07
0.20
0.10
0.02
0.05
4.89 Q1°99
2.41
0
-
135
TABLE 19. - continued
t
Storm
n
equation
r2
a
b
D50,
23 Sept.
24 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
7 Oct.
16 Oct.
4
5
6
18
5
8
12
14
7*1010
6.80
Qo.59
21.36 Q_03
7.70
oio
14.26 Q11
2*10° Q078
31.26 Q311
1949.2 0o.6
8.73 Q
0.08
0.06
0.07
0.02
0.40
0,03
0.12
0.09
*
No equations were developed for events represented by less than
three data points.
2Streamlow, Q, is
m3s1icn2
indicates significance at the 90% level of probability;
** indicates significance at the 95% level of probability.
136
TABLE 20.
Relationships1
Between Bedload Transport (BLD > 0.25
and BLD > 2.0 mm), Coarse Particulate Organic Matter
Transport (CPOM), Two Particle Diameters CD50 and D90),
and Streamf low2 for the Rising Limbs of Individual Storm
Events at the Upper Riffle of Trap Bay Creek, Chichagof
Island, Alaska, During the Fall, 19809 Storm Season
Storm
n
equation
BLD > 0.25
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
3
5
3
7
5
6
n,
24 Sept
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct
3
708
5
11.42
3
2570
7
687
5
27290
3175
6
,
092
054
3
3
7
Q149
Q256
Q014
Q357
2 88
Q
5
6
1039 Q147
59 Q6.055.
1303166
116
Q03
Q15
3 Q5
4227
3
5
3
7
5
6
0
Q
Q142
Q072
Q09
2 70
Q
__________
a
**
**
**
**
**
**
**
**
**
**
**
**
0.90
0.54
0.81
0.03
**
**
*
**
086
0.89
**
**
**
**
**
**
*
0.82
0.60
0.90
0.01
0.11
0.86
**
**
**
0.86
0.04
0.11
0.09
0.23
0.87
**
**
**
**.
*
*
=
-2.0
003 Q020
3.94
46.24
2.62
0.62
37.06
3
kghr1
1.36
CPOM, kg.hr
0.82
5
F
0
Q
D50,
24 Sept.
28 Sept
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
0.99
0.44
0.99
Q50
BLD > 20
2
kg.hr1
2.23
11220 Q128
1316 Q191
1288 Q01
1175
174582 Q0
1641
r
137
TABLE 20. - continued
Storm
n
equation
TI
90'
24 Sept.
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
3
5
=
-1.1
0.46 Q_03
9.27 Q086
3
66.07
7
15.24 Q092
31.41 Q007
5
6
O 46
15.45Q
r2
F3
a
b
0.86
0.05
0.54
0.05
0.12
0.02
No equations were developed for events represented by less than
three data points.
2Streamflow, Q, is in
m3si2,
indicates significance at the 90Z level of probability;
** indicates significance at the 95% level of probability.
138
TABLE 21.
Relationships1 Between Bedload Transport (BLD > 0.25 = and
BLD > 2.0
),Coarse Particulate Organic Matter Transport
(CPOM), Two Particle Diameters (D50 and D90), and Streaznflow2 for the Rising Limbs of Individual Stori Events at
the Lower Riffle of Trap Bay Creek, Chichagof Island,
A1aska During the Fa1l 1980, Stori Season
Stori
n
equation
BLD > 0.25
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
35 Q°
53 Q187
7
2646 Q4°14
66221
12647 Q
6
5
3
7
4
6
0.13
0.05
0.95
0.96
0.50
%5o
BLD > 2.0
28 Sept0
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
2
ti
F3
a
b
, kg.hr
3
4
r
**
**
**
**
**
**
**
**
039
**
**
**
*
*
**
*
kg.hr1
,
0 8
5 Q005
8 Q193
0.14
1829 Q506
0.97
0.99
112460
3 32
5834 Q
0
**
*
*
CPOM, kg.hr
28 Sept.
30 Sept.- 1 Oct.
5
189
0.59
**
3
152 Q233
1 Oct.
2 Oct.
5 Oct.
7
126
093
088
035
077
**
*
**
**
**
**
**
.
4
6
8
0.5 Q3 68
1178 Q
D50, =
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct.
5
0.19 Q'
0.37
3
2.10 Q0
0
7
4
6
427 Q271
6457
0 3
0.71 Q
0.10
0.29
0.03
**
139
TABLE 21. - continued
Storm
n
equation
r
2
t
P
3
a
b
D90, =
28 Sept.
30 Sept.- 1 Oct.
1 Oct.
2 Oct.
5 Oct0
5
3
7
4
6
1.15
36.78
17.38
392.64
-1.2
4571 Q
0.31
0.03
0.41
0.55
0.70
**
*
*
**
**
**
No equations were developed for events represented by less than
three data points.
2
3-1 -2
Streamf low, Q, is in m s
Icu
indicates significance at the 90% level of probability;
** indicates sigiiificance at the 95% level of probability.
140
Relationships1 between bedload transport (BLD > 0.25 = and
TA3LE 22.
BLD > 2.0 mm), Coarse Particulate Organic Matter Transport
(CPOM), Two Particle Diameters (D50 and D90), and Streanflow2 for the Falling Limbs of Individual StQrm Events a.t the Upper Riffle of Trap Bay Creeks Chichagof Island,
A1aska During the Fall, 1980, Storm Season
-
t-3
Storm
n
equation
, kg.hr
1.65
1479 Q
-0.5
824 Q
52722 Q751
BLD > 0,25
30
Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
3
9
4
6
7
926
3
9
1115
460
1.82
Q07
Q900
6
1116863 Q11
2.7*109 Q
7
516
4
0.68
0.15
0.15
0.51
0.48
Q
mm, kghr
Q190
F3
a
b
1
25*10
BLD > 2.0
30 Sept.- 1 Oct0
1 Oct.
5 Oct.
7 Oct.
16 Oct.
r2
**
**
*
**
**
*
**
1
*
0.53
017
0.17
0.70
0.59
**
**
**
**
**
**
*
**
**
CPOM, kg.hr
30. Sept.- 1 Oct.
3
1 Oct.
5 Oct.
7 Oct.
16 Oct,
9
4
6
7
56
63 Q2
96 Q
01 Q082
25 Q
D50
30 Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
3
9
4
6
7
4073
2.36
119.90
2269.90
2.43
0.98
0.20
0.26
0.,21
006
**
=
0.45
Q04
Q445
4.10
o 22
Q
0.13
0.20
0.26
0.21
0.06
**
141
TABLE 22. - continued
Storm
n
equatioii
r2
F3
a
0.03
0.23
0.88
0.21
0.09
**
**
**
b
D90, =1
30 Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
3
-0.2
7.77 Q_o.3
9
13.40 Q472
4
6
1202.28
17100.15 Q_0,5
7
7.76Q
*
**
1No equations were developed for events represented by less than
three data points.
2
Streamflow, Q, is in
3-1
s
-2
1
iiidicates significance at the 90% level of probability;
** indicates significance at the 95% level of probability.
142
TABLE 23.
Relationships1 Between Bedload Trausport (BLD > 0.25
and
BLD > 2.0 uim), Coarse Particulate Organic Matter Trausport
(CPOM), Two Particle Diameters (D50 and D90), and Streamflow2 for the Falling Limbs of Individual Storm Events at
the Lower Riffle of Trap Bay Creek, Chichagof Island,
Alaska, During the Fall, 1980, Storm Season
Storm
n
equation
BLD > 0.26
30
Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
3
9
4
6
7
, kg.hr
Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
3
9
4
6
7
2553
783610 Q
504 Q
30 Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct0
16 Oct.
3
,
b
a
*
0.23
0.32
**
**
**
*
**
*
**
**
**
**
kg.hr1
017
3.7*L05
1729 Q62
301995
0.01 Q1
290 Q
2
0.95
0.83
0.12
0.25
**
*
0.95
0.47
0.72
0,03
**
**
**
kg.hr1
2.7*1018Q461
Q70
9
4
103
3133 Q
6
1.9*104 Q,.
18 Q"°
7
t3
F3
0.36
0.97
092
009 Q16
CPOM,
2
1
4.7*106 Q94
BLD > 0.2
30
r
'
**
*
**
*
**
**
*
**
0
D50, =
30 Sept.- 1 Oct.
3
1 Oct.
5 Oct.
7 Oct.
9
9.12
4.11
4
0.92 Q04
0.01
0.76
0.25
6
l6Oct.
7
064 Q0 03
2.49Q
0
C
1
**
143
TABLE 23. - continued
Storm
n
equation
r2
P3
a
D90, =
30 Sept.- 1 Oct.
1 Oct.
5 Oct.
7 Oct.
16 Oct.
6.58 Q°
3
59.29
2.40
164.06 Q
11.05 Q
9
4
6
7
0.64
0.23
0.31
*
**
006
0.08
1No equations were develped fo* events represented by less than
three data points.
2
3-1 -2
Streamflow, Q, m s
1
indicates significance at the 90% level of probability;
** indicates significance at the 95% levl of probability.
b
APPENDIX 4.
I4orphometric Characteristics of Trap Bay Creek,
Chichagof Island, Alaska
144
TABLE 14.
Morphometric Characteristics of Trap Bay Creek, Chichagof
Is laud, Alaska
T'ap
Area
5.23 ml2
13.55
Perimeter - 9 ml - 14.5 1
Max. elev. = 3780 ft
0 ft
Mean elev.
1935 ft
Basin relief = 3400 ft
Relief ratio = 0.24
Miii. elev.
Total no. stres = 26
1 3rd-order, 5 2nd-order, 20 1st-order
area of
basins
mean
1st-order
2nd-order
3.0
34
0.15
0.68
.2
3rd-order
5 23 ml2
5.23
7.2 ml
length
2.1 mi
4.8 ml
mean
0.36
0.96
2.1
total stream length = 14.1 ml = 22.7 I
relief
mean
20650 ft
1033
5000 ft
1000
2510 ft
2510
Drain density =
2.7 mimI2
Constant of channel
maintenance = 1/DD =
0.37 mi2/mi
Stream frequency =
4.97
j2
Relative density =
0.68
SF/DD2
Mainstream slope = relief/length - 0.17 ft/ft Length to basin center
Mean slope 1st-order
0.54 ft/ft
- 1.63 mi
Mean slope 2nd-order = 0.20 ft/ft
Basin length = 2.69 ii
Basin width = 2.75 mi
Bifurcation ratio
4.47
Basin area ratio = 0.17
Stream length ratio = 0.41
Stre
relief ratio = 0.64
Stream slope ratio = 1.78
SLR/B = 10.78
(B)
(BAR)
(SLR)
(SRR)
(SRR)
Length of flow (5280/2*DD)= 978 ft
Lemniscate (BL2/4-A) = 0.35
Basin elongation ((A/3.14)½ - 2/BL)
= 0.96
Compactness coefficient = 1.11
Circularity = 0.81
Form Factor 1 (A/BL2) = 0.72
Form Factor 2 (BL/BW) = 0.98
Texture-Slope Product (DD*RR) = 0.65
Watershed Topography Factor (B*SSR/SLR) = 6.92